{"schema_version":"onlylabs.public_analysis_evidence.v1","title":"Google (DeepMind / Gemini) analysis evidence pack","description":"Public onlylabs evidence pack for cited agent analysis: captured pages, ranked public signals, and stored web-search provenance used by the background analysis workflow.","url":"https://onlylabs.fyi/analysis/google-deepmind","json_url":"https://onlylabs.fyi/analysis/google-deepmind/evidence.json","generated_at":"2026-06-13T15:16:44.832Z","org":{"slug":"google-deepmind","name":"Google (DeepMind / Gemini)","category":"frontier-lab","category_label":"Frontier lab","dossier_url":"https://onlylabs.fyi/labs/google-deepmind"},"analysis":{"url":"https://onlylabs.fyi/analysis/google-deepmind","json_url":"https://onlylabs.fyi/analysis/google-deepmind/analysis.json","generated_at":"2026-06-08T15:59:08.768+00:00"},"workflow":{"version":"synthesize-analyses","provider":null,"model":null,"agent":null,"public_pack_mode":"local-pages-and-events","live_web_fetches":false,"note":"Public evidence exports do not trigger live Exa calls; stored Exa provenance is included when analysis metadata contains it."},"stats":{"pages":28,"events":140,"web":0,"evidence":88,"signal_desks":{"hiring":12,"forks":0,"releases":32,"talking":13,"repos":3},"data_radar_lanes":{"data":1,"evals":1,"infrastructure":2,"safety":1,"product":0},"data_radar_matches":5,"stored_analysis_evidence":null,"stored_analysis_web":null,"stored_analysis_signal_desks":null,"stored_analysis_data_radar_lanes":null,"stored_analysis_data_radar_matches":null},"stored_web_provenance":null,"evidence":[{"ref":"P1","kind":"page","title":"google-deepmind/clrs repository metadata","date":"2026-06-13T08:03:33.888898+00:00","date_source":null,"source_url":"https://github.com/google-deepmind/clrs","signal_url":null,"signal_json_url":null,"text":"# google-deepmind/clrs\n\nLanguage: Jupyter Notebook\n\nLicense: Apache-2.0\n\nStars: 534\n\nForks: 116\n\nOpen issues: 11\n\nCreated: 2021-08-26T13:47:37Z\n\nPushed: 2026-06-12T19:44:26Z\n\nDefault branch: master\n\nFork: no\n\nArchived: no\n\nREADME:\n# The CLRS Algorithmic Reasoning Benchmark\n\nLearning representations of algorithms is an emerging area of machine learning,\nseeking to bridge concepts from neural networks with classical algorithms. The\nCLRS Algorithmic Reasoning Benchmark (CLRS) consolidates and extends previous\nwork toward evaluation algorithmic reasoning by providing a suite of\nimplementations of classical algorithms. These algorithms have been selected\nfrom the third edition of the standard *Introduction to Algorithms* by Cormen,\nLeiserson, Rivest and Stein.\n\n## Getting started\n\nThe CLRS Algorithmic Reasoning Benchmark can be installed with pip, either from\nPyPI:\n\n```shell\npip install dm-clrs\n```\n\nor directly from GitHub (updated more frequently):\n\n```shell\npip install git+https://github.com/google-deepmind/clrs.git\n```\n\nYou may prefer to install it in a virtual environment if any requirements\nclash with your Python installation:\n\n```shell\npython3 -m venv clrs_env\nsource clrs_env/bin/activate\npip install git+https://github.com/google-deepmind/clrs.git\n```\n\nOnce installed you can run our example baseline model:\n\n```shell\npython3 -m clrs.examples.run\n```\n\nIf this is the first run of the example, the dataset will be downloaded and\nstored in `--dataset_path` (default '/tmp/CLRS30').\nAlternatively, you can also download and extract https://storage.googleapis.com/dm-clrs/CLRS30_v1.0.0.tar.gz\n\n## Algorithms as graphs\n\nCLRS implements the selected algorithms in an idiomatic way, which aligns as\nclosely as possible to the original CLRS 3ed pseudocode. By controlling the\ninput data distribution to conform to the preconditions we are able to\nautomatically generate input/output pairs. We additionally provide trajectories\nof \"hints\" that expose the internal state of each algorithm, to both optionally\nsimplify the learning challenge and to distinguish between different algorithms\nthat solve the same overall task (e.g. sorting).\n\nIn the most generic sense, algorithms can be seen "},{"ref":"P2","kind":"page","title":"google-deepmind/clrs v0.0.2","date":"2026-06-13T08:03:33.218081+00:00","date_source":null,"source_url":"https://github.com/google-deepmind/clrs/releases/tag/v0.0.2","signal_url":null,"signal_json_url":null,"text":"# CLRS 0.0.2\n\nRepository: google-deepmind/clrs\n\nTag: v0.0.2\n\nPublished: 2021-08-26T17:40:38Z\n\nPrerelease: no\n\nRelease notes:\nThe CLRS Algorithmic Reasoning Benchmark."},{"ref":"P3","kind":"page","title":"google-deepmind/clrs v1.0.0","date":"2026-06-13T08:03:33.205845+00:00","date_source":null,"source_url":"https://github.com/google-deepmind/clrs/releases/tag/v1.0.0","signal_url":null,"signal_json_url":null,"text":"# CLRS 1.0.0\n\nRepository: google-deepmind/clrs\n\nTag: v1.0.0\n\nPublished: 2022-06-01T16:07:23Z\n\nPrerelease: no\n\nRelease notes:\n## Main changes\n\n* Extended the benchmark from 21 to 30 tasks by adding the following:\n- Activity selection (Gavril, 1972)\n- Longest common subsequence\n- Articulation points\n- Bridges\n- Kosaraju's strongly connected components algorithm (Aho et al., 1974)\n- Kruskal's minimum spanning tree algorithm (Kruskal, 1956)\n- Segment intersection\n- Graham scan convex hull algorithm (Graham, 1972)\n- Jarvis' march convex hull algorithm (Jarvis, 1973)\n* Added new baseline processors:\n- Deep Sets (Zaheer et al., NIPS 2017) and Pointer Graph Networks (Veličković et al., NeurIPS 2020) as particularisations of the existing Message-Passing Neural Network processor.\n- End-to-End Memory Networks (Sukhbaatar et al., NIPS 2015)\n- Graph Attention Networks v2 (Brody et al., ICLR 2022)\n\n## Detailed changes\n\n* Add PyPI installation instructions. by @copybara-service in https://github.com/deepmind/clrs/pull/6\n* Fix README typo. by @copybara-service in https://github.com/deepmind/clrs/pull/7\n* Expose `Sampler` base class in public API. by @copybara-service in https://github.com/deepmind/clrs/pull/8\n* Add dataset reader. by @copybara-service in https://github.com/deepmind/clrs/pull/12\n* Patch imbalanced samplers for DFS-based algorithms. by @copybara-service in https://github.com/deepmind/clrs/pull/15\n* Disk-based samplers for convex hull algorithms. by @copybara-service in https://github.com/deepmind/clrs/pull/16\n* Avoid dividing by zero in F_1 score computaton. by @copybara-service in https://github.com/deepmind/clrs/pull/18\n* Sparsify the graphs generated for Kruskal. by @copybara-service in https://github.com/deepmind/clrs/pull/20\n* Option to add an lstm after the processor. by @copybara-service in https://github.com/deepmind/clrs/pull/19\n* Include dataset class and creation using tensorflow_datasets format. by @copybara-service in https://github.com/deepmind/clrs/pull/23\n* Change types of DataPoint and DataPoint members. by @copybara-service in https://github.com/deepmind/clrs/pull/22\n* Remove unnecessary data loading procedures. by @copybara-service in https://g"},{"ref":"P4","kind":"page","title":"google-deepmind/clrs v2.0.3","date":"2026-06-13T08:03:33.10793+00:00","date_source":null,"source_url":"https://github.com/google-deepmind/clrs/releases/tag/v2.0.3","signal_url":null,"signal_json_url":null,"text":"# v2.0.3\n\nRepository: google-deepmind/clrs\n\nTag: v2.0.3\n\nPublished: 2025-01-25T22:43:20Z\n\nPrerelease: no\n\nRelease notes:\n## What's Changed\n* Increment minor version number for 2.0.3 release. by @copybara-service in https://github.com/google-deepmind/clrs/pull/164\n\n**Full Changelog**: https://github.com/google-deepmind/clrs/compare/v2.0.2...v2.0.3"},{"ref":"P5","kind":"page","title":"google-deepmind/clrs v2.0.2","date":"2026-06-13T08:03:33.104397+00:00","date_source":null,"source_url":"https://github.com/google-deepmind/clrs/releases/tag/v2.0.2","signal_url":null,"signal_json_url":null,"text":"# v2.0.2\n\nRepository: google-deepmind/clrs\n\nTag: v2.0.2\n\nPublished: 2025-01-25T19:16:49Z\n\nPrerelease: no\n\nRelease notes:\n## What's Changed\n* Update version numbers from requirements.txt by @copybara-service in https://github.com/google-deepmind/clrs/pull/156\n* Log the warning only for the first on-the-fly sampler. by @copybara-service in https://github.com/google-deepmind/clrs/pull/157\n* Update run.py. by @copybara-service in https://github.com/google-deepmind/clrs/pull/158\n* Resolve unsoundness caught by pytype --strict-none-binding. by @copybara-service in https://github.com/google-deepmind/clrs/pull/161\n\n**Full Changelog**: https://github.com/google-deepmind/clrs/compare/v2.0.1...v2.0.2"},{"ref":"P6","kind":"page","title":"google-deepmind/clrs v2.0.1","date":"2026-06-13T08:03:33.072219+00:00","date_source":null,"source_url":"https://github.com/google-deepmind/clrs/releases/tag/v2.0.1","signal_url":null,"signal_json_url":null,"text":"# v2.0.1 CLRS-Text Algorithmic Reasoning Language Benchmark\n\nRepository: google-deepmind/clrs\n\nTag: v2.0.1\n\nPublished: 2024-07-18T19:03:48Z\n\nPrerelease: no\n\nRelease notes:\n## Main Changes\n**Added the CLRS-Text Algorithmic Reasoning Language Benchmark to the codebase**\n\nCLRS-Text is a textual version of the traces generated by thirty algorithms selected from the third edition of the standard \"Introduction to Algorithms\" textbook by Cormen, Leiserson, Rivest, and Stein. It serves to consolidate and unify previous lines of research in this direction and offers a robust test bed for evaluating language models' out-of-distribution reasoning capabilities.\n\nCLRS-Text is Hugging Face compatible and can generate data in many formats, such as JSON.\n\nFor more details, refer to the [[README.md](https://github.com/google-deepmind/clrs/blob/master/clrs/_src/clrs_text/README.md)].\n\n**Base CLRS script update**\nRevamped the base script incorporating modeling and data improvements from the [A Generalist Neural Algorithmic Learner](https://arxiv.org/pdf/2209.11142) paper.\n\n## What's Changed\n* Add `ml_collections` to `requirements.txt`. by @copybara-service in https://github.com/google-deepmind/clrs/pull/139\n* Added script to generate json for all algorithms. by @copybara-service in https://github.com/google-deepmind/clrs/pull/140\n* Adding huggingface generators for clrs text by @mcleish7 in https://github.com/google-deepmind/clrs/pull/135\n* Add clrs_text to __init__.py by @copybara-service in https://github.com/google-deepmind/clrs/pull/144\n* Roll forward PR #104 by @RerRayne in https://github.com/google-deepmind/clrs/pull/146\n* Add CLRS-Text details in the README files. by @copybara-service in https://github.com/google-deepmind/clrs/pull/148\n\n## New Contributors\n* @sigeisler made their first contribution in https://github.com/google-deepmind/clrs/pull/104\n* @mcleish7 made their first contribution in https://github.com/google-deepmind/clrs/pull/135\n* @RerRayne made their first contribution in https://github.com/google-deepmind/clrs/pull/146\n* @gurux13 made their first contribution in https://github.com/google-deepmind/clrs/pull/147\n\n**Full Changelog**: https://github.com/google-de"},{"ref":"P7","kind":"page","title":"google-deepmind/chex v0.1.92","date":"2026-06-13T08:03:32.785865+00:00","date_source":null,"source_url":"https://github.com/google-deepmind/chex/releases/tag/v0.1.92","signal_url":null,"signal_json_url":null,"text":"# Chex 0.1.92\n\nRepository: google-deepmind/chex\n\nTag: v0.1.92\n\nPublished: 2026-06-12T14:27:45Z\n\nPrerelease: no\n\nRelease notes:\n## What's Changed\n* update docs builder by @copybara-service[bot] in https://github.com/google-deepmind/chex/pull/410\n* [pmap] Update logic that works with both PmapSharding and NamedSharding. by @copybara-service[bot] in https://github.com/google-deepmind/chex/pull/414\n* [pmap] Avoid degraded performance under the new `jax.pmap`. by @copybara-service[bot] in https://github.com/google-deepmind/chex/pull/419\n* Fix a deprecation warning for new versions of JAX and Numpy. Fixes #424. by @copybara-service[bot] in https://github.com/google-deepmind/chex/pull/425\n* Add strict mode support in chex.assert_trees_all_close. This aligns the behaviour with chex.assert_trees_all_equal. by @copybara-service[bot] in https://github.com/google-deepmind/chex/pull/436\n* [pmap] Remove `jax.config.pmap_shmap_merge`. by @copybara-service[bot] in https://github.com/google-deepmind/chex/pull/438\n* [pmap] In-line definitions of `jax.device_put_sharded` and `jax.device_put_replicated`. by @copybara-service[bot] in https://github.com/google-deepmind/chex/pull/440\n* fix tests and drop python 3.11 by @copybara-service[bot] in https://github.com/google-deepmind/chex/pull/442\n* Release v0.1.92 by @copybara-service[bot] in https://github.com/google-deepmind/chex/pull/443\n\n**Full Changelog**: https://github.com/google-deepmind/chex/compare/v0.1.91...v0.1.92"},{"ref":"P8","kind":"page","title":"google-deepmind/rlax v0.1.9","date":"2026-06-13T08:03:32.76188+00:00","date_source":null,"source_url":"https://github.com/google-deepmind/rlax/releases/tag/v0.1.9","signal_url":null,"signal_json_url":null,"text":"# RLax 0.1.9\n\nRepository: google-deepmind/rlax\n\nTag: v0.1.9\n\nPublished: 2026-06-12T23:37:49Z\n\nPrerelease: no\n\nRelease notes:\n## What's Changed\n* Support of mask for learnable steps in v-trace implementation. by @copybara-service[bot] in https://github.com/google-deepmind/rlax/pull/151\n* fix tests and drop python 3.11 by @copybara-service[bot] in https://github.com/google-deepmind/rlax/pull/163\n* Release v0.1.9 by @copybara-service[bot] in https://github.com/google-deepmind/rlax/pull/164\n\n**Full Changelog**: https://github.com/google-deepmind/rlax/compare/v0.1.8...v0.1.9"},{"ref":"P9","kind":"page","title":"google-deepmind/distrax v0.1.9","date":"2026-06-13T08:03:32.635462+00:00","date_source":null,"source_url":"https://github.com/google-deepmind/distrax/releases/tag/v0.1.9","signal_url":null,"signal_json_url":null,"text":"# Distrax 0.1.9\n\nRepository: google-deepmind/distrax\n\nTag: v0.1.9\n\nPublished: 2026-06-12T23:38:54Z\n\nPrerelease: no\n\nRelease notes:\n## What's Changed\n* Avoid usage of deprecated `jax.core` APIs. by @copybara-service[bot] in https://github.com/google-deepmind/distrax/pull/326\n* [JAX] Replace uses of jnp.float_, jnp.int_, and jnp.complex_ with float, int, and complex in .astype, jnp.zeros, and jnp.ones. by @copybara-service[bot] in https://github.com/google-deepmind/distrax/pull/327\n* fix tests and drop python 3.11 by @copybara-service[bot] in https://github.com/google-deepmind/distrax/pull/332\n* Release v0.1.9 by @copybara-service[bot] in https://github.com/google-deepmind/distrax/pull/333\n\n**Full Changelog**: https://github.com/google-deepmind/distrax/compare/v0.1.8...v0.1.9"},{"ref":"P10","kind":"page","title":"google-deepmind/mctx repository metadata","date":"2026-06-11T20:05:13.881998+00:00","date_source":null,"source_url":"https://github.com/google-deepmind/mctx","signal_url":null,"signal_json_url":null,"text":"# google-deepmind/mctx\n\nDescription: Monte Carlo tree search in JAX\n\nLanguage: Python\n\nLicense: Apache-2.0\n\nStars: 2631\n\nForks: 209\n\nOpen issues: 7\n\nCreated: 2022-03-01T17:26:43Z\n\nPushed: 2026-06-10T14:02:08Z\n\nDefault branch: main\n\nFork: no\n\nArchived: no\n\nREADME:\n# Mctx: MCTS-in-JAX\n\nMctx is a library with a [JAX](https://github.com/google/jax)-native\nimplementation of Monte Carlo tree search (MCTS) algorithms such as\n[AlphaZero](https://deepmind.com/blog/article/alphazero-shedding-new-light-grand-games-chess-shogi-and-go),\n[MuZero](https://deepmind.com/blog/article/muzero-mastering-go-chess-shogi-and-atari-without-rules), and\n[Gumbel MuZero](https://openreview.net/forum?id=bERaNdoegnO). For computation\nspeed up, the implementation fully supports JIT-compilation. Search algorithms\nin Mctx are defined for and operate on batches of inputs, in parallel. This\nallows to make the most of the accelerators and enables the algorithms to work\nwith large learned environment models parameterized by deep neural networks.\n\n## Installation\n\nYou can install the latest released version of Mctx from PyPI via:\n\n```sh\npip install mctx\n```\n\nor you can install the latest development version from GitHub:\n\n```sh\npip install git+https://github.com/google-deepmind/mctx.git\n```\n\n## Motivation\n\nLearning and search have been important topics since the early days of AI\nresearch. In the [words of Rich Sutton](http://www.incompleteideas.net/IncIdeas/BitterLesson.html):\n\n> One thing that should be learned [...] is the great power of general purpose\n> methods, of methods that continue to scale with increased computation even as\n> the available computation becomes very great. The two methods that seem to\n> scale arbitrarily in this way are *search* and *learning*.\n\nRecently, search algorithms have been successfully combined with learned models\nparameterized by deep neural networks, resulting in some of the most powerful\nand general reinforcement learning algorithms to date (e.g. MuZero).\nHowever, using search algorithms in combination with deep neural networks\nrequires efficient implementations, typically written in fast compiled\nlanguages; this can come at the expense of usability and hackabilit"},{"ref":"P11","kind":"page","title":"google-deepmind/fancyflags repository metadata","date":"2026-06-11T20:05:13.876921+00:00","date_source":null,"source_url":"https://github.com/google-deepmind/fancyflags","signal_url":null,"signal_json_url":null,"text":"# google-deepmind/fancyflags\n\nDescription: A Python library for defining structured command-line flags.\n\nLanguage: Python\n\nLicense: Apache-2.0\n\nStars: 35\n\nForks: 7\n\nOpen issues: 1\n\nCreated: 2021-02-05T09:14:53Z\n\nPushed: 2026-06-10T22:18:46Z\n\nDefault branch: master\n\nFork: no\n\nArchived: no\n\nREADME:\n# fancyflags\n\n<!--* freshness: { owner: 'ydoron' reviewed: '2025-02-07' } *-->\n\n![PyPI Python version](https://img.shields.io/pypi/pyversions/fancyflags)\n![PyPI version](https://badge.fury.io/py/fancyflags.svg)\n\n## Introduction\n\n`fancyflags` is a Python library that extends\n[`absl.flags`](https://github.com/abseil/abseil-py) with additional structured\nflag types.\n\n`fancyflags` provides flags corresponding to structures such as\n[dicts](#dict-flags), [dataclasses, and (somewhat) arbitrary callables](#auto).\n\nThese flags are typed and validated like regular `absl` flags, catching errors\nbefore they propagate into your program. To override values, users can access\nfamiliar \"dotted\" flag names.\n\nTIP: Already a `fancyflags` user? Check out our [usage tips](#tips)!\n\n## A short note on design philosophy:\n\n`fancyflags` promotes mixing with regular `absl` flags. In many cases a few\nregular `absl` flags are all you need!\n\n`fancyflags` does not require you to modify library code: it should only be used\nin your \"main\" file\n\n`fancyflags` is not a dependency injection framework, and avoids\nprogramming-language-like power features. We prefer that users write regular\nPython for wiring up their code, because it's explicit, simple to understand,\nand allows static analysis tools to identify problems.\n\n## Installation\n\n`fancyflags` can be installed from PyPI using `pip`:\n\n```shell\npip install fancyflags\n```\n\nIt can also be installed directly from our GitHub repository:\n\n```shell\npip install git+git://github.com/deepmind/fancyflags.git\n```\n\nor alternatively by checking out a local copy of our repository and running:\n\n```shell\npip install /path/to/local/fancyflags/\n```\n\n## Dict flags\n\nIf we have a class `Replay`, with arguments `capacity`, `priority_exponent` and\nothers, we can define a corresponding dict flag in our main script\n\n```python\nimport fancyflags as ff\n\n_REPLAY_FLAG = ff.DEFINE_dic"},{"ref":"P12","kind":"page","title":"google-deepmind/mctx v0.0.2","date":"2026-06-11T20:05:13.176132+00:00","date_source":null,"source_url":"https://github.com/google-deepmind/mctx/releases/tag/v0.0.2","signal_url":null,"signal_json_url":null,"text":"# mctx 0.0.2\n\nRepository: google-deepmind/mctx\n\nTag: v0.0.2\n\nPublished: 2022-07-26T10:18:52Z\n\nPrerelease: no\n\nRelease notes:\n## What's Changed\n* Add a link to a0-jax on Connect Four, Gomoku, and Go.\n* Improve debugging by avoiding an unused NaN.\n* Use jax.tree_util to avoid deprecation warnings.\n\n**Full Changelog**: https://github.com/deepmind/mctx/commits/v0.0.2"},{"ref":"P13","kind":"page","title":"google-deepmind/mctx v0.0.3","date":"2026-06-11T20:05:13.059432+00:00","date_source":null,"source_url":"https://github.com/google-deepmind/mctx/releases/tag/v0.0.3","signal_url":null,"signal_json_url":null,"text":"# mctx 0.0.3\n\nRepository: google-deepmind/mctx\n\nTag: v0.0.3\n\nPublished: 2023-07-24T20:29:08Z\n\nPrerelease: no\n\nRelease notes:\n## What's Changed\n* Add experimental support for Stochastic MuZero.\n* Fix pylint and pytype warnings.\n\n**Full Changelog**: https://github.com/deepmind/mctx/compare/v0.0.2...v0.0.3"},{"ref":"P14","kind":"page","title":"google-deepmind/mctx v0.0.4","date":"2026-06-11T20:05:12.969025+00:00","date_source":null,"source_url":"https://github.com/google-deepmind/mctx/releases/tag/v0.0.4","signal_url":null,"signal_json_url":null,"text":"# mctx 0.0.4\n\nRepository: google-deepmind/mctx\n\nTag: v0.0.4\n\nPublished: 2023-11-24T10:49:59Z\n\nPrerelease: no\n\nRelease notes:\n## What's Changed\n* Automatically determine num_actions and num_chance_outcomes in stochastic_muzero_policy (thanks Carlos Martin).\n* Explicitly use int32 for the argmax output even when using jax_enable_x64.\n\n**Full Changelog**: https://github.com/google-deepmind/mctx/compare/v0.0.3...v0.0.4"},{"ref":"P15","kind":"page","title":"google-deepmind/mctx v0.0.5","date":"2026-06-11T20:05:12.940383+00:00","date_source":null,"source_url":"https://github.com/google-deepmind/mctx/releases/tag/v0.0.5","signal_url":null,"signal_json_url":null,"text":"# mctx 0.0.5\n\nRepository: google-deepmind/mctx\n\nTag: v0.0.5\n\nPublished: 2023-11-24T11:52:10Z\n\nPrerelease: no\n\nRelease notes:\n## What's Changed\n* Automatically determine num_actions and num_chance_outcomes in stochastic_muzero_policy (thanks Carlos Martin).\n* Explicitly use int32 for the argmax output even when using jax_enable_x64.\n\n**Full Changelog**: https://github.com/google-deepmind/mctx/compare/v0.0.3...v0.0.5"},{"ref":"P16","kind":"page","title":"google-deepmind/mctx v0.0.6","date":"2026-06-11T20:05:12.689435+00:00","date_source":null,"source_url":"https://github.com/google-deepmind/mctx/releases/tag/v0.0.6","signal_url":null,"signal_json_url":null,"text":"# Mctx 0.0.6\n\nRepository: google-deepmind/mctx\n\nTag: v0.0.6\n\nPublished: 2025-09-02T09:58:49Z\n\nPrerelease: no\n\nRelease notes:\n## What's Changed\n* Add a link to mctx-az. by @copybara-service[bot] in https://github.com/google-deepmind/mctx/pull/90\n* Replace deprecated `jax.tree_*` functions with `jax.tree.*` by @copybara-service[bot] in https://github.com/google-deepmind/mctx/pull/91\n* Fix typo in tree.py by @KazukiOhta in https://github.com/google-deepmind/mctx/pull/92\n* Change deprecated jax.tree_util.tree_map to jax.tree.map. Fix argument passed to jax.numpy.finfo call. by @carlosgmartin in https://github.com/google-deepmind/mctx/pull/95\n* Ignore pylint refactoring messages. by @copybara-service[bot] in https://github.com/google-deepmind/mctx/pull/101\n\n## New Contributors\n* @KazukiOhta made their first contribution in https://github.com/google-deepmind/mctx/pull/92\n\n**Full Changelog**: https://github.com/google-deepmind/mctx/compare/v0.0.5...v0.0.6"},{"ref":"P17","kind":"page","title":"DiffusionGemma: 4x faster text generation","date":"2026-06-11T20:04:59.238547+00:00","date_source":null,"source_url":"https://deepmind.google/blog/diffusiongemma-4x-faster-text-generation/","signal_url":null,"signal_json_url":null,"text":"Introducing DiffusionGemma \n\nDiffusionGemma: 4x faster text generation\n\nJun 10, 2026\n\n· \n\nShare\n\nx.com\n\nFacebook\n\nLinkedIn\n\nMail\n\nCopy link\n\nOur newest open experimental model delivers up to 4x faster inference on dedicated GPUs and opens the door to exploring speed-critical, interactive local workflows.\n\nBrendan O&#x27;Donoghue\n\nResearch Scientist\n\nSebastian Flennerhag\n\nResearch Scientist\n\nShare\n\nx.com\n\nFacebook\n\nLinkedIn\n\nMail\n\nCopy link\n\nToday, we’re introducing DiffusionGemma, an experimental open model that explores text diffusion, an exceptionally fast approach to text generation. Released under an Apache 2.0 license, this 26B Mixture of Experts (MoE) model moves beyond the sequential token-by-token processing of typical autoregressive Large Language Models (LLMs). Instead, it generates entire blocks of text simultaneously, delivering up to 4x faster text generation on GPUs.\n\nBuilt upon the industry-leading intelligence-per-parameter of our Gemma 4 family and cutting-edge Gemini Diffusion research , DiffusionGemma integrates a novel diffusion head designed to maximize generation speed. While autoregressive Gemma 4 models remain the standard for high-quality production outputs, DiffusionGemma is designed for researchers and developers exploring speed-critical, interactive local workflows such as in-line editing, rapid iteration, and generating non-linear text structures.\nUnlocking new value for developers\nDevelopers building real-time interactive AI applications often struggle with the latency bottlenecks of local inference. DiffusionGemma addresses these challenges directly, with some key trade-offs:\nBlazing fast inference: By shifting the decode bottleneck from memory-bandwidth to compute, DiffusionGemma generates up to 4x faster token output on dedicated GPUs. (1000+ tokens per second on a single NVIDIA H100, 700+ tokens per second on NVIDIA GeForce RTX 5090). \n\n1 \n\nAccessible hardware footprint: Operating as a 26B total Mixture of Experts (MoE) model that activates only 3.8B parameters during inference, DiffusionGemma fits comfortably within 18GB VRAM limits of high-end dedicated consumer GPUs when quantized.\nBi-directional attention : Generating 256 to"},{"ref":"P18","kind":"page","title":"google/diffusiongemma-26B-A4B-it model card","date":"2026-06-11T20:04:55.883317+00:00","date_source":null,"source_url":"https://huggingface.co/google/diffusiongemma-26B-A4B-it/raw/main/README.md","signal_url":null,"signal_json_url":null,"text":"---\nlicense: apache-2.0\nlicense_link: https://ai.google.dev/gemma/docs/gemma_4_license\npipeline_tag: image-text-to-text\nlibrary_name: transformers\n---\n\n<div align=\"center\">\n<img src=https://ai.google.dev/gemma/images/diffusiongemma_banner.png>\n</div>\n\n<p align=\"center\">\n<a href=\"https://huggingface.co/google/diffusiongemma-26B-A4B-it\" target=\"_blank\">Hugging Face</a> |\n<a href=\"https://github.com/google-gemma\" target=\"_blank\">GitHub</a> |\n<a href=\"https://blog.google/innovation-and-ai/technology/developers-tools/diffusion-gemma-faster-text-generation/\" target=\"_blank\">Launch Blog</a> |\n<a href=\"https://ai.google.dev/gemma/docs/diffusiongemma\" target=\"_blank\">Documentation</a>\n<br>\n<b>License</b>: <a href=\"https://ai.google.dev/gemma/docs/gemma_4_license\" target=\"_blank\">Apache 2.0</a> | <b>Authors</b>: <a href=\"https://deepmind.google/models/gemma/\" target=\"_blank\">Google DeepMind</a>\n</p>\n\nDiffusionGemma is a generative model built by Google DeepMind. Based on the 26B A4B Mixture-of-Experts (MoE) Gemma 4 architecture, DiffusionGemma generates tokens using discrete diffusion. This open-weights model is multimodal, handling text, image, and video inputs to generate text output.\n\nBuilt on a MoE foundation, DiffusionGemma is designed to improve generation speed (tokens per second) while remaining deployable across various hardware environments. DiffusionGemma builds upon the architectural and capability advancements of Gemma 4, introducing several core features:\n\n* **Discrete Text Diffusion** – Shifts from token-by-token autoregression to block-autoregressive multi-canvas sampling. It generates text by iteratively denoising blocks of tokens (a 'canvas') in parallel, significantly increasing decoding speed. \n* **Multimodal Input Processing** – Processes interleaved text, image (with variable aspect ratio and resolution support), and video inputs to generate text outputs. \n* **Encoder-Decoder Architecture** – Utilizes an autoregressive encoder to process and cache the prompt context, paired with a decoder that applies bidirectional attention over the generation canvas. \n* **Mixture-of-Experts (MoE) Efficiency** – Leverages a sparse MoE design (8 active experts out of"},{"ref":"P19","kind":"page","title":"Investing in multi-agent AI safety research","date":"2026-06-11T20:04:46.927687+00:00","date_source":null,"source_url":"https://deepmind.google/blog/investing-in-multi-agent-ai-safety-research/","signal_url":null,"signal_json_url":null,"text":"Google DeepMind and partners announce multi-agent safety research funding call. — Google DeepMind Skip to main content \n\nJune 11, 2026 Responsibility & Safety \nInvesting in multi-agent AI safety research\nGoogle DeepMind, Schmidt Sciences, the Cooperative AI Foundation, ARIA and Google.org\n\nShare\n\nScaling AI Safety Research for a Multi-Agent World\nFor the past decade, we’ve focused on making individual AI models more capable, helpful and safe. Today, Google DeepMind — together with Schmidt Sciences , the Cooperative AI Foundation , the Advanced Research and Invention Agency , and supported by Google.org — is announcing a new technical research funding call of up to $10M for researchers worldwide.\nAs AI technology scales, we’re entering a new era. Soon, millions of AI agents — built by different organizations — will interact across digital environments, communicating, negotiating and transacting with one another.\nWhen these systems interact, they must do so safely and predictably. This shift creates a vital opportunity: we can strengthen the safety and stability of the entire AI ecosystem from the very beginning.\nThe funding call focuses on the study of how large-scale multi-agent AI systems behave as a group, and how we can provide frameworks to understand and mitigate against potential risks. By empowering researchers globally, we aim to solve the “invisible” safety risks that arise when independent systems interact across different networks.\nWhy the agent ecosystem matters\nWhen large groups of AI agents interact, new collective behaviors and capabilities can emerge suddenly. Currently, we lack the tools to predict, measure and monitor these transitions. Most safety evaluations analyze models in isolation. However, as we and others have previously argued, interacting autonomous agents can produce complex, \"emergent\" behaviors that are difficult to anticipate.\nBecause this is a new area of research, it is critical to understand how these shifts occur. For example, could they cause an unpredictable flurry of economic activity or lead to new security challenges? Understanding how to manage these system-wide behaviors is our core objective.\nScaling the frontier of m"},{"ref":"P20","kind":"page","title":"google-deepmind/llms_can_learn_rules repository metadata","date":"2026-06-11T04:00:40.745622+00:00","date_source":null,"source_url":"https://github.com/google-deepmind/llms_can_learn_rules","signal_url":null,"signal_json_url":null,"text":"# google-deepmind/llms_can_learn_rules\n\nLanguage: Python\n\nLicense: Apache-2.0\n\nStars: 63\n\nForks: 12\n\nOpen issues: 1\n\nCreated: 2024-12-06T01:51:35Z\n\nPushed: 2024-12-06T01:54:21Z\n\nDefault branch: main\n\nFork: no\n\nArchived: no\n\nREADME:\n# llms_can_learn_rules\n\n![HtT prompting](asset/htt.svg)\n\nA major reason for failure of chain-of-thought prompting is its tendency to\nhallucinate rules in multi-step reasoning. Our work, Hypotheses-to-Theories (HtT),\nprompts LLMs to induce rules from training samples, build a rule library, and\napply it to solve reasoning problems.\n\nHtT can be viewed as a new paradigm of learning with LLMs. Instead of learning\nmodel parameters, HtT learns a rule library that transfers across different models\nand textual forms. It can be applied to black-box LLMs such as GPT-4 and Gemini-Pro.\n\n## Installation\n\nFirst, let's create a virtual environment:\n\nBelow we will use conda environment. If you haven't heard about conda, you can install the miniconda following https://docs.conda.io/projects/miniconda/en/latest/\n\n```bash\nconda create -n htt python=3.10\nconda activate htt\n```\n\nInstall dependencies through pip.\n\n```bash\npip install -r requirements.txt\n```\n\nIf it is your first time using nltk, please also download the necessary pieces:\n\n```python\n>>> import nltk\n>>> nltk.download('punkt')\n>>> nltk.download('punkt_tab')\n```\n\nExport your API keys as an environment variable.\n\n```bash\nexport OPENAI_API_KEY=your-open-ai-key\nexport GOOGLE_API_KEY=your-google-key\n```\n\n## 🛠️ Usage ##\n\n### Dataset ###\n\nPlease navigate to `artifacts/dataset` and run the `download.sh` script to prepare the datasets.\n\nUse `source/train.py` for training HtT and `source/test.py` for testing HtT and\nother methods. All the experiments will be logged in `experiment/`.\n\n### Training (Induction Stage) ###\n\nBelow are the command lines for training HtT. Note if you train HtT with GPT-4,\nit will incur substantial time (2-21 hours) and cost ($20-270 USD) per dataset.\nIf you only want to reproduce the results of HtT, we highly recommend you to use\nthe provided checkpoints of rule libraries.\n\n```bash\npython source/train.py -c artifacts/config/clutrr/symbolic/gpt-4_5-shot_cot_htt.yaml --num-iterat"},{"ref":"P21","kind":"page","title":"google-deepmind/cube repository metadata","date":"2026-06-11T04:00:39.463905+00:00","date_source":null,"source_url":"https://github.com/google-deepmind/cube","signal_url":null,"signal_json_url":null,"text":"# google-deepmind/cube\n\nDescription: Code for CUBE extraction and cultural diversity metric introduced in \"Beyond Aesthetics: Cultural Competence in Text-to-Image Models\"\n\nLanguage: Python\n\nLicense: Apache-2.0\n\nStars: 7\n\nForks: 0\n\nOpen issues: 0\n\nCreated: 2025-01-18T18:12:27Z\n\nPushed: 2025-01-20T11:45:11Z\n\nDefault branch: main\n\nFork: no\n\nArchived: no\n\nREADME:\n## Code for [Beyond Aesthetics: Cultural Competence in Text-to-Image Models](https://arxiv.org/abs/2407.06863) published in NeurIPS 2024 (Track on Datasets and Benchmarks)\n\nCurrent T2I benchmarks primarily focus on faithfulness, aesthetics, and realism of generated images, overlooking the critical dimension of cultural competence. In this work, we introduce a framework to evaluate cultural competence of T2I models along two crucial dimensions: cultural awareness and cultural diversity, and present a scalable approach using a combination of structured knowledge base (KB) and large language models (LLMs) to build a large dataset of cultural artifacts to enable this evaluation. In particular, we apply this approach to build CUBE (CUltural BEnchmark for Text-to-Image models), a first-of-its-kind benchmark to evaluate cultural competence of T2I models. CUBE covers cultural artifacts associated with 8 countries across different geo-cultural regions and along 3 concepts: cuisine, landmarks, and art. We also introduce cultural diversity as a novel T2I evaluation component, leveraging quality-weighted Vendi score. In this repository, we release the 1) code for CUBE extraction from WikiData, 2) a notebook [![Open In\nColab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/google-deepmind/cube/blob/master/cultural_diversity.ipynb) with the implementation of the cultural diversity evaluation aspect for T2I and 3) CUBE dataset containing CUBE CSpace and CUBE-1K prompts.\n\n## CUBE Extraction\n\n### Setup\n\n#### Fetching latest KB dump (redoing this will refresh latest Wikidata dump)\n\n```\npython3.8 -m venv wikidata\nsource wikidata/bin/activate\npip3 install https://ringgaard.com/data/dist/sling-3.0.0-py3-none-linux_x86_64.whl\npip install urllib3\npip install absl-py\npip install "},{"ref":"P22","kind":"page","title":"google-deepmind/proactive_t2i_agents repository metadata","date":"2026-06-11T04:00:39.353361+00:00","date_source":null,"source_url":"https://github.com/google-deepmind/proactive_t2i_agents","signal_url":null,"signal_json_url":null,"text":"# google-deepmind/proactive_t2i_agents\n\nDescription: Code release for the paper, \"Proactive Agents for Text-to-Image Generation under Uncertainty\"\n\nLanguage: Python\n\nLicense: Apache-2.0\n\nStars: 71\n\nForks: 7\n\nOpen issues: 1\n\nCreated: 2024-12-07T05:35:11Z\n\nPushed: 2025-07-28T19:01:11Z\n\nDefault branch: main\n\nFork: no\n\nArchived: no\n\nREADME:\n# [Proactive Agents for Multi-Turn Text-to-Image Generation under Uncertainty](https://arxiv.org/abs/2412.06771)\n\nUser prompts for generative AI models are often underspecified or open-ended,\nwhich may lead to sub-optimal responses. This prompt underspecification problem\nis particularly evident in text-to-image (T2I) generation, where users commonly\nstruggle to articulate their precise intent. This disconnect between the user’s vision\nand the model’s interpretation often forces users to painstakingly and repeatedly\nrefine their prompts. To address this, we propose a design for proactive T2I agents\nequipped with an interface to actively ask clarification questions when uncertain,\nand present their understanding of user intent as an interpretable belief graph\nthat a user can edit. We build simple prototypes for such agents and verify their\neffectiveness through both human studies and automated evaluation. We observed\nthat at least 90% of human subjects found these agents and their belief graphs\nhelpful for their T2I workflow. Moreover, we use a scalable automated evaluation\napproach using two agents, one with a ground truth image and the other tries to\nask as few questions as possible to align with the ground truth. On DesignBench, a\nbenchmark we created for artists and designers, the COCO dataset (Lin et al., 2014)\nand ImageInWords ([Garg et al., 2024](https://arxiv.org/abs/2405.02793)), we observed that these T2I agents were able\nto ask informative questions and elicit crucial information to achieve successful\nalignment with at least 2 times higher VQAScore ([Lin et al., 2024](https://arxiv.org/abs/2404.01291)) than the standard\nsingle-turn T2I generation. Play with our agent demo [![Open In\nColab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/google-deepmind/proactive_t2i_age"},{"ref":"P23","kind":"page","title":"google-deepmind/physics-IQ-benchmark repository metadata","date":"2026-06-11T04:00:39.13066+00:00","date_source":null,"source_url":"https://github.com/google-deepmind/physics-IQ-benchmark","signal_url":null,"signal_json_url":null,"text":"# google-deepmind/physics-IQ-benchmark\n\nDescription: Benchmarking physical understanding in generative video models\n\nLanguage: Python\n\nLicense: NOASSERTION\n\nStars: 302\n\nForks: 33\n\nOpen issues: 3\n\nCreated: 2024-12-14T05:07:19Z\n\nPushed: 2026-05-27T13:13:57Z\n\nDefault branch: main\n\nFork: no\n\nArchived: no\n\nREADME:\n<p align=\"center\">\n<img src=\"assets/duckie.png\" width=\"20%\" alt=\"logo\">\n</p>\n\n[Step A: Generating Videos](#step-a-generating-videos-for-physics-iq-test-cases-based-on-video-model) | [Step B: Evaluating Generated Videos](#step-b-evaluating-generated-videos-on-physics-iq-to-generate-benchmark-scores) | [Leaderboard](#leaderboard) | [Citation](#citation) | [License](#license-and-disclaimer)\n\n# Physics-IQ: Benchmarking physical understanding in generative video models\n\nPhysics-IQ is a high-quality, realistic, and comprehensive benchmark dataset for evaluating physical understanding in generative video models.\n\nProject website: [physics-iq.github.io](https://physics-iq.github.io/)\n\n### Key Features:\n- **Real-world videos**: All videos are captured with high-quality cameras, not rendered.\n- **Diverse scenarios**: Covers a wide range of physical phenomena, including collisions, fluid dynamics, gravity, material properties, light, shadows, magnetism, and more.\n- **Multiple perspectives**: Each scenario is filmed from 3 different angles.\n- **Variations**: Each scenario is recorded twice to capture natural physical variations.\n- **High resolution and frame rate**: Videos are recorded at 3840 × 2160 resolution and 30 frames per second.\n\n<p align=\"center\">\n<img src=\"assets/teaser1.gif\" width=\"23%\" alt=\"Teaser 1\">\n<img src=\"assets/teaser2.gif\" width=\"23%\" alt=\"Teaser 2\">\n<img src=\"assets/teaser3.gif\" width=\"23%\" alt=\"Teaser 3\">\n<img src=\"assets/teaser4.gif\" width=\"23%\" alt=\"Teaser 4\">\n<img src=\"assets/teaser5.gif\" width=\"23%\" alt=\"Teaser 5\">\n<img src=\"assets/teaser6.gif\" width=\"23%\" alt=\"Teaser 6\">\n<img src=\"assets/teaser7.gif\" width=\"23%\" alt=\"Teaser 7\">\n<img src=\"assets/teaser8.gif\" width=\"23%\" alt=\"Teaser 8\">\n</p>\n\n---\n\n## Leaderboard\nThe best possible score on Physics-IQ is 100.0%, this score would be achieved by physically realistic videos that differ only in physi"},{"ref":"P24","kind":"page","title":"google-deepmind/lm_act repository metadata","date":"2026-06-11T04:00:39.076311+00:00","date_source":null,"source_url":"https://github.com/google-deepmind/lm_act","signal_url":null,"signal_json_url":null,"text":"# google-deepmind/lm_act\n\nDescription: LMAct: A Benchmark for In-Context Imitation Learning with Long Multimodal Demonstrations\n\nLanguage: Python\n\nLicense: Apache-2.0\n\nStars: 31\n\nForks: 6\n\nOpen issues: 0\n\nCreated: 2024-12-06T22:32:54Z\n\nPushed: 2025-05-21T12:59:14Z\n\nDefault branch: main\n\nFork: no\n\nArchived: no\n\nREADME:\n# LMAct: A Benchmark for In-Context Imitation Learning with Long Multimodal Demonstrations\n\n<p align=\"center\">\n<img src=\"https://raw.githubusercontent.com/google-deepmind/lm_act/master/overview.svg\" alt=\"Overview figure\"/>\n</p>\n\nThis repository provides an implementation of our ICML 2025 paper [LMAct: A Benchmark for In-Context Imitation Learning with Long Multimodal Demonstrations](https://arxiv.org/abs/2412.01441).\n\n> In this paper, we present a benchmark to pressure-test today’s frontier\nmodels’ multimodal decision-making capabilities in the very long-context regime\n(up to one million tokens) and investigate whether these models can learn from\nlarge numbers of expert demonstrations in their context.\nWe evaluate the performance of Claude 3.5 Sonnet, Gemini 1.5 Flash, Gemini 1.5\nPro, Gemini 2.0 Flash Experimental, GPT-4o, o1-mini, o1-preview, and o1 as\npolicies across a battery of simple interactive decision-making tasks: playing\ntic-tac-toe, chess, and Atari, navigating grid worlds, solving crosswords, and\ncontrolling a simulated cheetah.\nWe study increasing amounts of expert demonstrations in the context — from no\ndemonstrations to 512 full episodes.\nAcross our tasks, models rarely manage to fully reach expert performance, and\noften, presenting more demonstrations has little effect.\nSome models steadily improve with more demonstrations on a few tasks.\nWe investigate the effect of encoding observations as text or images and the\nimpact of chain-of-thought prompting.\nTo help quantify the impact of other approaches and future innovations, we open\nsource our benchmark that covers the zero-, few-, and many-shot regimes in a\nunified evaluation.\n\n## Contents\n\n```\n.\n|\n├── crafter - Crafter (needs to be downloaded)\n|\n├── data - Expert demonstrations (need to be downloaded)\n|\n├── src\n| ├── agents\n| │   ├── chess.py - Stockfish agent (chess expert)\n| │   ├─"},{"ref":"P25","kind":"page","title":"google-deepmind/md4 repository metadata","date":"2026-06-11T04:00:38.999218+00:00","date_source":null,"source_url":"https://github.com/google-deepmind/md4","signal_url":null,"signal_json_url":null,"text":"# google-deepmind/md4\n\nDescription: Official Jax Implementation of MD4 Masked Diffusion Models\n\nLanguage: Python\n\nLicense: Apache-2.0\n\nStars: 160\n\nForks: 17\n\nOpen issues: 2\n\nCreated: 2024-12-09T18:34:11Z\n\nPushed: 2025-02-27T17:10:20Z\n\nDefault branch: main\n\nFork: no\n\nArchived: no\n\nREADME:\n# MD4: Simplified and Generalized Masked Diffusion for Discrete Data\n\n<img src=\"imgs/imagenet64.gif\" width=\"600\" />\n\n## Installation\n\n### Create Python environment\n\nplease use `requirements_gpu.txt` if your accelerator is GPUs, use\n`requirements_tpu`.txt when using Google Cloud TPUs.\n\n```\npython -m venv md4_venv\nsource md4_venv/bin/activate\npip install -r requirements_[gpu/tpu].txt\nexport PYTHONPATH=\"$PYTHONPATH:~/path/to/md4\"\n```\n\n## Usage\n\nprepare openwebtext for training (i.e., tokenize and pack examples)\n\n```\nmkdir data_dir\npython prepare_openwebtext_data.py\n```\n\ntrain a MD4-S model over text data (OpenWebText, Fineweb-EDU).\n\n```\npython md4/main.py --config=md4/configs/md4/openwebtext.py --sharded=false --workdir=./expt\n```\n\nalternatively, you can train a MD4-S model over image data (CIFAR-10,\nImageNet-64).\n\n```\npython md4/main.py --config=md4/configs/md4/cifar10.py --sharded=false --workdir=./expt\n```\n\n### choose batch size\n\nBatch size depends on your compute resource. For training a MD4-S model with\nsequence length 1024, eight `A100` GPUs can support a maximum batch size of\n`128`. If running on TPUs, eight `v5litepod` chips can support a maximum batch\nsize of `32`.\n\n## Citing this work\n\nAdd citation details here, usually a pastable BibTeX snippet:\n\n```\n@inproceedings{shi2024simplified,\ntitle={Simplified and Generalized Masked Diffusion for Discrete Data},\nauthor={Shi, Jiaxin and Han, Kehang and Wang, Zhe and Doucet, Arnaud and Titsias, Michalis K.},\nbooktitle={Advances in Neural Information Processing Systems},\nyear={2024}\n}\n```\n\n## License and disclaimer\n\nCopyright 2024 DeepMind Technologies Limited\n\nAll software is licensed under the Apache License, Version 2.0 (Apache 2.0);\nyou may not use this file except in compliance with the Apache 2.0 license.\nYou may obtain a copy of the Apache 2.0 license at:\nhttps://www.apache.org/licenses/LICENSE-2.0\n\nUnless required by applica"},{"ref":"P26","kind":"page","title":"google-deepmind/librispeech-long repository metadata","date":"2026-06-11T04:00:38.256316+00:00","date_source":null,"source_url":"https://github.com/google-deepmind/librispeech-long","signal_url":null,"signal_json_url":null,"text":"# google-deepmind/librispeech-long\n\nDescription: LibriSpeech-Long is a benchmark dataset for long-form speech generation and processing. Released as part of \"Long-Form Speech Generation with Spoken Language Models\" (arXiv 2024).\n\nLicense: CC-BY-4.0\n\nStars: 98\n\nForks: 4\n\nOpen issues: 1\n\nCreated: 2024-12-21T00:50:08Z\n\nPushed: 2024-12-28T18:42:46Z\n\nDefault branch: main\n\nFork: no\n\nArchived: no\n\nREADME:\n# LibriSpeech-Long\n\n**[Download audio, ground-truth transcripts, and per-file durations for all splits (3GB).](https://storage.googleapis.com/librispeech_long/v0_1.tar.gz)** \n\nThis is a benchmark dataset for evaluating long-form variants of speech processing tasks such as speech continuation, speech recognition, and text-to-speech synthesis. It is derived from the [LibriSpeech](https://www.openslr.org/12) dev and test sets, whose utterances are reprocessed into contiguous examples of up to 4 minutes in length (in the manner of [LibriLight's `cut_by_vad.py` script](https://github.com/facebookresearch/libri-light/blob/main/data_preparation/cut_by_vad.py)).\n\n**For more details:** see \"Long-Form Speech Generation with Spoken Language Models\" ([paper](https://arxiv.org/abs/2412.18603), [website](https://google.github.io/tacotron/publications/speechssm/)).\n\n*This is part of a preprint that is work-in-progress; dataset may be subject to change.*\n\n## Citation\n\nWhen using this dataset, please cite the associated [paper](https://arxiv.org/abs/2412.18603):\n\n```\n@article{park2024long,\nauthor = {Se Jin Park and\nJulian Salazar and\nAren Jansen and\nKeisuke Kinoshita and\nYong Man Ro and\nR. J. Skerry{-}Ryan},\ntitle = {Long-Form Speech Generation with Spoken Language Models},\njournal = {CoRR},\nvolume = {abs/2412.18603},\nyear = {2024}\n}\n```\n\n## License and disclaimer\n\nCopyright 2024 DeepMind Technologies Limited\n\nThe software and materials, except for the underlying LibriSpeech data, are licensed under the Creative Commons Attribution 4.0 International License (CC-BY). You may obtain a copy of the CC-BY license at: https://creativecommons.org/licenses/by/4.0/legalcode, or in the LICENSE file.\n\nThe materials contain adapted material from the LibriSpeech dataset. LibriSpeech is also licens"},{"ref":"P27","kind":"page","title":"google-deepmind/inli repository metadata","date":"2026-06-11T04:00:38.164218+00:00","date_source":null,"source_url":"https://github.com/google-deepmind/inli","signal_url":null,"signal_json_url":null,"text":"# google-deepmind/inli\n\nLicense: CC-BY-SA-4.0\n\nStars: 5\n\nForks: 1\n\nOpen issues: 0\n\nCreated: 2025-01-09T22:38:48Z\n\nPushed: 2025-01-23T04:16:58Z\n\nDefault branch: main\n\nFork: no\n\nArchived: no\n\nREADME:\n# Implied NLI\n\nCode & resources for the paper [Entailed Between the Lines: Incorporating Implication into NLI](https://arxiv.org/abs/2501.07719)\n\n### Dataset Overview\n**Implied NLI**, or **INLI** is a corpus of 10k premises that mirror real-world communication and 40k hypotheses that are implied, explicit, neutral, and contradictory.\n\n* Number of examples: 40,000\n* Number of labels: 4 (implied entailment, explicit entailment, neutral, contradictory)\n\nOn top of the raw data, we also include all prompts used to generate INLI, all human annotations described in the paper, and the outputs of the LLM experiments described in the paper.\n\nFor more details on the design and content of the dataset, please see [our paper](https://arxiv.org/abs/2501.07719)\n\n### Data Format\nOur raw dataset is split into three csv files. \n\n* Size of training dataset: 32,000\n* Size of test dataset: 4,000\n* Size of validation dataset: 4,000\n\nIn each of these files, one row represents a premise and four corresponding hypotheses. These files include the following columns:\n\n* `dataset`: The original dataset containing the conversation or social norm used to create the premise\n* `premise`: A premise\n* `implied_entailment`: A hypothesis implicitly entailed in the premise\n* `explicit_entailment`: A hypothesis explicitly entailed in the premise\n* `neutral`: A hypothesis neither entailed nor contradicted by the premise\n* `contradiction`: A hypothesis contradicted by the premise\n\n## Citation\n```\n@misc{havaldar2025entailedlinesincorporatingimplication,\ntitle={Entailed Between the Lines: Incorporating Implication into NLI}, \nauthor={Shreya Havaldar and Hamidreza Alvari and John Palowitch and Mohammad Javad Hosseini and Senaka Buthpitiya and Alex Fabrikant},\nyear={2025},\neprint={2501.07719},\narchivePrefix={arXiv},\nprimaryClass={cs.CL},\nurl={https://arxiv.org/abs/2501.07719}, \n}\n```\n\n## Contact\n[Shreya Havaldar](https://shreyahavaldar.com/)\n\n## Disclaimer\nINLI may contain premises and hypotheses involving sensit"},{"ref":"P28","kind":"page","title":"google-deepmind/sustainable_behavior repository metadata","date":"2026-06-11T04:00:38.124814+00:00","date_source":null,"source_url":"https://github.com/google-deepmind/sustainable_behavior","signal_url":null,"signal_json_url":null,"text":"# google-deepmind/sustainable_behavior\n\nLanguage: Jupyter Notebook\n\nLicense: Apache-2.0\n\nStars: 15\n\nForks: 4\n\nOpen issues: 0\n\nCreated: 2024-12-31T03:57:50Z\n\nPushed: 2025-01-23T20:16:28Z\n\nDefault branch: main\n\nFork: no\n\nArchived: no\n\nREADME:\n# sustainable_behavior\n\nThis repository contains the code accompanying the publication\n**Deep reinforcement learning can promote sustainable human\nbehaviour in a common-pool resource problem**,\npublished in the Nature Communications, 2025.\n\n## Usage\n\n### Figures\n\nYou can load the dataset and the code used to produce the figures in the\nresearch article using Colab:\n[![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/google-deepmind/sustainable_behavior/blob/main/notebooks/sustainable_behavior.ipynb)\n\nClone this repository with\n\n```sh\ngit clone https://github.com/google-deepmind/sustainable_behavior.git\n```\n\n## Citing this work\n\n```latex\n@article{kosterpislar2025sustainable,\ntitle={Deep reinforcement learning can promote sustainable human\nbehaviour in a common-pool resource problem},\nauthor={Koster, R AND\nPislar, M AND\nTacchetti, A AND\nBalaguer, J AND\nLiu, L AND\nElie, R AND\nHauser, OP AND\nTuyls, K AND\nBotvinick, M AND\nSummerfield, C},\njournal={Nature Communications},\nvolume={XXX},\nnumber={XX},\npages={X},\ndoi={XXX},\nURL={XXX},\npublisher={Nature Publishing Group UK London}\nyear={2025},\n}\n```\n\n## License and disclaimer\n\nCopyright 2024 DeepMind Technologies Limited\n\nAll software is licensed under the Apache License, Version 2.0 (Apache 2.0);\nyou may not use this file except in compliance with the Apache 2.0 license.\nYou may obtain a copy of the Apache 2.0 license at:\nhttps://www.apache.org/licenses/LICENSE-2.0\n\nAll other materials are licensed under the Creative Commons Attribution 4.0\nInternational License (CC-BY). You may obtain a copy of the CC-BY license at:\nhttps://creativecommons.org/licenses/by/4.0/legalcode\n\nUnless required by applicable law or agreed to in writing, all software and\nmaterials distributed here under the Apache 2.0 or CC-BY licenses are\ndistributed on an \"AS IS\" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND,\neither express or implied. See"},{"ref":"E1","kind":"event","title":"AlphaEvolve: How our Gemini-powered coding agent is scaling impact across fields","date":"2026-05-07T00:00:00.000Z","date_source":"page.visible_date","source_url":"https://deepmind.google/blog/alphaevolve-impact/","signal_url":"https://onlylabs.fyi/signals/5cc4cd52-fe73-44d3-9adf-1337fa2de766","signal_json_url":"https://onlylabs.fyi/signals/5cc4cd52-fe73-44d3-9adf-1337fa2de766/signal.json","text":"post_published · AlphaEvolve: How our Gemini-powered coding agent is scaling impact across fields · signal_desk=talking · occurred_at=2026-05-07T00:00:00.000Z · url=https://deepmind.google/blog/alphaevolve-impact/ · hn=327 points/149 comments · data_radar_lanes=Infrastructure · data_radar_terms=infra, infrastructure, scaling · data_radar_reason=Google (DeepMind / Gemini) has a writing signal matching infrastructure. · raw={\"excerpt\":\"Explore how AlphaEvolve's Gemini-powered algorithms are driving impact across business, infrastructure, and science.\"}"},{"ref":"E2","kind":"event","title":"Reimagining the mouse pointer for the AI era","date":"2026-05-12T00:00:00.000Z","date_source":"page.visible_date","source_url":"https://deepmind.google/blog/ai-pointer/","signal_url":"https://onlylabs.fyi/signals/88ab0c15-fe35-4383-9581-2d598488b11a","signal_json_url":"https://onlylabs.fyi/signals/88ab0c15-fe35-4383-9581-2d598488b11a/signal.json","text":"post_published · Reimagining the mouse pointer for the AI era · signal_desk=talking · occurred_at=2026-05-12T00:00:00.000Z · url=https://deepmind.google/blog/ai-pointer/ · hn=252 points/213 comments · raw={\"excerpt\":\"Google DeepMind is transforming the mouse pointer into a context-aware AI partner. Move beyond the friction of traditional prompting with intuitive AI collaboration in Chrome and beyond.\"}"},{"ref":"E3","kind":"event","title":"Gemini Robotics-ER 1.6: Powering real-world robotics tasks through enhanced embodied reasoning","date":"2026-04-14T00:00:00.000Z","date_source":"page.visible_date","source_url":"https://deepmind.google/blog/gemini-robotics-er-1-6/","signal_url":"https://onlylabs.fyi/signals/d03c05ad-4033-4f78-a620-ebe0f20fc0a4","signal_json_url":"https://onlylabs.fyi/signals/d03c05ad-4033-4f78-a620-ebe0f20fc0a4/signal.json","text":"post_published · Gemini Robotics-ER 1.6: Powering real-world robotics tasks through enhanced embodied reasoning · signal_desk=talking · occurred_at=2026-04-14T00:00:00.000Z · url=https://deepmind.google/blog/gemini-robotics-er-1-6/ · hn=219 points/84 comments · raw={\"excerpt\":\"Gemini Robotics ER 1.6: Enhancing spatial reasoning and multi-view understanding for autonomous robotics.\"}"},{"ref":"E4","kind":"event","title":"google/gemma-4-12B-it","date":"2026-05-23T01:17:15+00:00","date_source":"source","source_url":"https://huggingface.co/google/gemma-4-12B-it","signal_url":"https://onlylabs.fyi/signals/92fc70be-efa5-4056-bcf1-d4a7a773fbe1","signal_json_url":"https://onlylabs.fyi/signals/92fc70be-efa5-4056-bcf1-d4a7a773fbe1/signal.json","text":"model_released · google/gemma-4-12B-it · signal_desk=releases · occurred_at=2026-05-23T01:17:15+00:00 · url=https://huggingface.co/google/gemma-4-12B-it · hf_downloads=911544 · hf_likes=977 · hf_params=11959730224 · pipeline=any-to-any · license=apache-2.0 · raw={\"derived_reason\":\"first-party-finetune\"}"},{"ref":"E5","kind":"event","title":"google/diffusiongemma-26B-A4B-it","date":"2026-06-09T12:40:12+00:00","date_source":"source","source_url":"https://huggingface.co/google/diffusiongemma-26B-A4B-it","signal_url":"https://onlylabs.fyi/signals/dc3f7316-5b55-4076-b61a-9a8ffd7a39de","signal_json_url":"https://onlylabs.fyi/signals/dc3f7316-5b55-4076-b61a-9a8ffd7a39de/signal.json","text":"model_released · google/diffusiongemma-26B-A4B-it · signal_desk=releases · occurred_at=2026-06-09T12:40:12+00:00 · url=https://huggingface.co/google/diffusiongemma-26B-A4B-it · hf_downloads=20669 · hf_likes=639 · hf_params=25823778864 · pipeline=image-text-to-text · license=apache-2.0"},{"ref":"E6","kind":"event","title":"google/gemma-4-12B","date":"2026-05-23T02:43:54+00:00","date_source":"source","source_url":"https://huggingface.co/google/gemma-4-12B","signal_url":"https://onlylabs.fyi/signals/77187ecb-8ac8-47a1-9bcb-17c5ed9128d2","signal_json_url":"https://onlylabs.fyi/signals/77187ecb-8ac8-47a1-9bcb-17c5ed9128d2/signal.json","text":"model_released · google/gemma-4-12B · signal_desk=releases · occurred_at=2026-05-23T02:43:54+00:00 · url=https://huggingface.co/google/gemma-4-12B · hf_downloads=198271 · hf_likes=527 · hf_params=11959730224 · pipeline=any-to-any · license=apache-2.0"},{"ref":"E7","kind":"event","title":"google/gemma-4-31B-it-assistant","date":"2026-04-23T18:55:29+00:00","date_source":"source","source_url":"https://huggingface.co/google/gemma-4-31B-it-assistant","signal_url":"https://onlylabs.fyi/signals/f5f13d8a-21a6-42c3-9430-4ed4fdb8ba2a","signal_json_url":"https://onlylabs.fyi/signals/f5f13d8a-21a6-42c3-9430-4ed4fdb8ba2a/signal.json","text":"model_released · google/gemma-4-31B-it-assistant · signal_desk=releases · occurred_at=2026-04-23T18:55:29+00:00 · url=https://huggingface.co/google/gemma-4-31B-it-assistant · hf_downloads=298736 · hf_likes=301 · hf_params=469518596 · pipeline=any-to-any · license=apache-2.0"},{"ref":"E8","kind":"event","title":"Decoupled DiLoCo: A new frontier for resilient, distributed AI training","date":"2026-04-23T00:00:00.000Z","date_source":"page.visible_date","source_url":"https://deepmind.google/blog/decoupled-diloco/","signal_url":"https://onlylabs.fyi/signals/52fc4c01-b93a-4c73-bb74-779e2884f255","signal_json_url":"https://onlylabs.fyi/signals/52fc4c01-b93a-4c73-bb74-779e2884f255/signal.json","text":"post_published · Decoupled DiLoCo: A new frontier for resilient, distributed AI training · signal_desk=talking · occurred_at=2026-04-23T00:00:00.000Z · url=https://deepmind.google/blog/decoupled-diloco/ · hn=49 points/6 comments · data_radar_lanes=Infrastructure · data_radar_terms=training · data_radar_reason=Google (DeepMind / Gemini) has a writing signal matching infrastructure. · raw={\"excerpt\":\"\"}"},{"ref":"E9","kind":"event","title":"google/magenta-realtime-2","date":"2026-05-28T21:55:47+00:00","date_source":"source","source_url":"https://huggingface.co/google/magenta-realtime-2","signal_url":"https://onlylabs.fyi/signals/d2518865-3640-4b31-b489-a6df8c03b63e","signal_json_url":"https://onlylabs.fyi/signals/d2518865-3640-4b31-b489-a6df8c03b63e/signal.json","text":"model_released · google/magenta-realtime-2 · signal_desk=releases · occurred_at=2026-05-28T21:55:47+00:00 · url=https://huggingface.co/google/magenta-realtime-2 · hf_downloads=6491 · hf_likes=185 · pipeline=text-to-audio · license=cc-by-4.0"},{"ref":"E10","kind":"event","title":"google/gemma-4-26B-A4B-it-assistant","date":"2026-04-23T20:16:46+00:00","date_source":"source","source_url":"https://huggingface.co/google/gemma-4-26B-A4B-it-assistant","signal_url":"https://onlylabs.fyi/signals/9e927ea2-d026-4b03-b840-197418c6a577","signal_json_url":"https://onlylabs.fyi/signals/9e927ea2-d026-4b03-b840-197418c6a577/signal.json","text":"model_released · google/gemma-4-26B-A4B-it-assistant · signal_desk=releases · occurred_at=2026-04-23T20:16:46+00:00 · url=https://huggingface.co/google/gemma-4-26B-A4B-it-assistant · hf_downloads=126521 · hf_likes=162 · hf_params=419711236 · pipeline=any-to-any · license=apache-2.0"},{"ref":"E11","kind":"event","title":"google/gemma-4-E4B-it-assistant","date":"2026-04-23T18:44:33+00:00","date_source":"source","source_url":"https://huggingface.co/google/gemma-4-E4B-it-assistant","signal_url":"https://onlylabs.fyi/signals/4d6ecb17-3a75-463d-996b-e52836554f99","signal_json_url":"https://onlylabs.fyi/signals/4d6ecb17-3a75-463d-996b-e52836554f99/signal.json","text":"model_released · google/gemma-4-E4B-it-assistant · signal_desk=releases · occurred_at=2026-04-23T18:44:33+00:00 · url=https://huggingface.co/google/gemma-4-E4B-it-assistant · hf_downloads=52245 · hf_likes=108 · hf_params=78779908 · pipeline=any-to-any · license=apache-2.0"},{"ref":"E12","kind":"event","title":"google/tipsv2-b14","date":"2026-04-09T06:28:59+00:00","date_source":"source","source_url":"https://huggingface.co/google/tipsv2-b14","signal_url":"https://onlylabs.fyi/signals/e2db2580-0b41-44b1-8747-d40f084a1b12","signal_json_url":"https://onlylabs.fyi/signals/e2db2580-0b41-44b1-8747-d40f084a1b12/signal.json","text":"model_released · google/tipsv2-b14 · signal_desk=releases · occurred_at=2026-04-09T06:28:59+00:00 · url=https://huggingface.co/google/tipsv2-b14 · hf_downloads=15201 · hf_likes=107 · hf_params=195948288 · pipeline=zero-shot-image-classification · license=apache-2.0"},{"ref":"E13","kind":"event","title":"google/gemma-4-12B-it-assistant","date":"2026-05-23T01:17:47+00:00","date_source":"source","source_url":"https://huggingface.co/google/gemma-4-12B-it-assistant","signal_url":"https://onlylabs.fyi/signals/17fdf1ca-1620-449e-ac95-34cbd68e26cc","signal_json_url":"https://onlylabs.fyi/signals/17fdf1ca-1620-449e-ac95-34cbd68e26cc/signal.json","text":"model_released · google/gemma-4-12B-it-assistant · signal_desk=releases · occurred_at=2026-05-23T01:17:47+00:00 · url=https://huggingface.co/google/gemma-4-12B-it-assistant · hf_downloads=29174 · hf_likes=82 · hf_params=422856964 · pipeline=any-to-any · license=apache-2.0"},{"ref":"E14","kind":"event","title":"google/gemma-4-12B-it-qat-q4_0-unquantized","date":"2026-06-04T05:20:28+00:00","date_source":"source","source_url":"https://huggingface.co/google/gemma-4-12B-it-qat-q4_0-unquantized","signal_url":"https://onlylabs.fyi/signals/84a69417-279f-40e1-8ebc-df721d885b16","signal_json_url":"https://onlylabs.fyi/signals/84a69417-279f-40e1-8ebc-df721d885b16/signal.json","text":"model_released · google/gemma-4-12B-it-qat-q4_0-unquantized · signal_desk=releases · occurred_at=2026-06-04T05:20:28+00:00 · url=https://huggingface.co/google/gemma-4-12B-it-qat-q4_0-unquantized · hf_downloads=17689 · hf_likes=44 · hf_params=11959730224 · pipeline=any-to-any · license=apache-2.0 · raw={\"derived_reason\":\"first-party-finetune\"}"},{"ref":"E15","kind":"event","title":"google/gemma-4-E2B-it-assistant","date":"2026-04-23T18:35:40+00:00","date_source":"source","source_url":"https://huggingface.co/google/gemma-4-E2B-it-assistant","signal_url":"https://onlylabs.fyi/signals/081c0046-2877-4281-a77c-f8add4460803","signal_json_url":"https://onlylabs.fyi/signals/081c0046-2877-4281-a77c-f8add4460803/signal.json","text":"model_released · google/gemma-4-E2B-it-assistant · signal_desk=releases · occurred_at=2026-04-23T18:35:40+00:00 · url=https://huggingface.co/google/gemma-4-E2B-it-assistant · hf_downloads=19215 · hf_likes=62 · hf_params=77993476 · pipeline=any-to-any · license=apache-2.0"},{"ref":"E16","kind":"event","title":"google/gemma-4-12B-it-qat-q4_0-unquantized-assistant","date":"2026-06-04T22:58:12+00:00","date_source":"source","source_url":"https://huggingface.co/google/gemma-4-12B-it-qat-q4_0-unquantized-assistant","signal_url":"https://onlylabs.fyi/signals/f58bdfb9-0e0c-4b1a-bb6d-8a3624fac327","signal_json_url":"https://onlylabs.fyi/signals/f58bdfb9-0e0c-4b1a-bb6d-8a3624fac327/signal.json","text":"model_released · google/gemma-4-12B-it-qat-q4_0-unquantized-assistant · signal_desk=releases · occurred_at=2026-06-04T22:58:12+00:00 · url=https://huggingface.co/google/gemma-4-12B-it-qat-q4_0-unquantized-assistant · hf_downloads=1873 · hf_likes=17 · hf_params=422856964 · pipeline=any-to-any · license=apache-2.0 · raw={\"derived_reason\":\"first-party-finetune\"}"},{"ref":"E17","kind":"event","title":"google/gemma-4-31B-it-qat-q4_0-unquantized-assistant","date":"2026-05-29T19:19:43+00:00","date_source":"source","source_url":"https://huggingface.co/google/gemma-4-31B-it-qat-q4_0-unquantized-assistant","signal_url":"https://onlylabs.fyi/signals/bdf07c85-5eb4-4573-b293-bc19df10e294","signal_json_url":"https://onlylabs.fyi/signals/bdf07c85-5eb4-4573-b293-bc19df10e294/signal.json","text":"model_released · google/gemma-4-31B-it-qat-q4_0-unquantized-assistant · signal_desk=releases · occurred_at=2026-05-29T19:19:43+00:00 · url=https://huggingface.co/google/gemma-4-31B-it-qat-q4_0-unquantized-assistant · hf_downloads=15996 · hf_likes=14 · hf_params=469518596 · pipeline=image-text-to-text · license=apache-2.0 · raw={\"derived_reason\":\"first-party-finetune\"}"},{"ref":"E18","kind":"event","title":"Strengthening Singapore’s AI Future: A New National Partnership","date":"2026-05-20T00:00:00.000Z","date_source":"page.visible_date","source_url":"https://deepmind.google/blog/strengthening-singapores-ai-future-a-new-national-partnership/","signal_url":"https://onlylabs.fyi/signals/776c96f8-a579-4ff9-8d39-ed1ed677ce69","signal_json_url":"https://onlylabs.fyi/signals/776c96f8-a579-4ff9-8d39-ed1ed677ce69/signal.json","text":"post_published · Strengthening Singapore’s AI Future: A New National Partnership · signal_desk=talking · occurred_at=2026-05-20T00:00:00.000Z · url=https://deepmind.google/blog/strengthening-singapores-ai-future-a-new-national-partnership/ · hn=3 points/0 comments · raw={\"excerpt\":\"Google DeepMind and Singapore partner to apply frontier AI to address complex challenges across health, education, and sustainability and more.\"}"},{"ref":"E19","kind":"event","title":"Co-Scientist: A multi-agent AI partner to accelerate research","date":"2026-05-19T00:00:00.000Z","date_source":"page.visible_date","source_url":"https://deepmind.google/blog/co-scientist-a-multi-agent-ai-partner-to-accelerate-research/","signal_url":"https://onlylabs.fyi/signals/59a5b24f-797e-444d-90f5-3bfb3db5e14f","signal_json_url":"https://onlylabs.fyi/signals/59a5b24f-797e-444d-90f5-3bfb3db5e14f/signal.json","text":"post_published · Co-Scientist: A multi-agent AI partner to accelerate research · signal_desk=talking · occurred_at=2026-05-19T00:00:00.000Z · url=https://deepmind.google/blog/co-scientist-a-multi-agent-ai-partner-to-accelerate-research/ · hn=3 points/0 comments · raw={\"excerpt\":\"Introducing Co-Scientist, a collaborative AI partner built with Gemini to help researchers accelerate scientific breakthroughs.\"}"},{"ref":"E20","kind":"event","title":"google/gemma-4-31B-it-qat-q4_0-unquantized","date":"2026-04-28T20:55:54+00:00","date_source":"source","source_url":"https://huggingface.co/google/gemma-4-31B-it-qat-q4_0-unquantized","signal_url":"https://onlylabs.fyi/signals/301aa2a7-d0cb-4954-9bbf-e8cc43149b90","signal_json_url":"https://onlylabs.fyi/signals/301aa2a7-d0cb-4954-9bbf-e8cc43149b90/signal.json","text":"model_released · google/gemma-4-31B-it-qat-q4_0-unquantized · signal_desk=releases · occurred_at=2026-04-28T20:55:54+00:00 · url=https://huggingface.co/google/gemma-4-31B-it-qat-q4_0-unquantized · hf_downloads=5600 · hf_likes=24 · hf_params=32682372656 · pipeline=image-text-to-text · license=apache-2.0 · raw={\"derived_reason\":\"first-party-finetune\"}"},{"ref":"E21","kind":"event","title":"google/gemma-4-E4B-it-qat-q4_0-unquantized-assistant","date":"2026-05-29T19:19:00+00:00","date_source":"source","source_url":"https://huggingface.co/google/gemma-4-E4B-it-qat-q4_0-unquantized-assistant","signal_url":"https://onlylabs.fyi/signals/a005e3c1-efd1-442c-9e6f-65ed49231271","signal_json_url":"https://onlylabs.fyi/signals/a005e3c1-efd1-442c-9e6f-65ed49231271/signal.json","text":"model_released · google/gemma-4-E4B-it-qat-q4_0-unquantized-assistant · signal_desk=releases · occurred_at=2026-05-29T19:19:00+00:00 · url=https://huggingface.co/google/gemma-4-E4B-it-qat-q4_0-unquantized-assistant · hf_downloads=881 · hf_likes=8 · hf_params=78779908 · pipeline=any-to-any · license=apache-2.0 · raw={\"derived_reason\":\"first-party-finetune\"}"},{"ref":"E22","kind":"event","title":"google/gemma-4-26B-A4B-it-qat-q4_0-unquantized-assistant","date":"2026-05-29T19:19:16+00:00","date_source":"source","source_url":"https://huggingface.co/google/gemma-4-26B-A4B-it-qat-q4_0-unquantized-assistant","signal_url":"https://onlylabs.fyi/signals/bc524648-11ff-4365-8b3e-a7793a0d2d4d","signal_json_url":"https://onlylabs.fyi/signals/bc524648-11ff-4365-8b3e-a7793a0d2d4d/signal.json","text":"model_released · google/gemma-4-26B-A4B-it-qat-q4_0-unquantized-assistant · signal_desk=releases · occurred_at=2026-05-29T19:19:16+00:00 · url=https://huggingface.co/google/gemma-4-26B-A4B-it-qat-q4_0-unquantized-assistant · hf_downloads=611 · hf_likes=8 · hf_params=419711236 · pipeline=image-text-to-text · license=apache-2.0 · raw={\"derived_reason\":\"first-party-finetune\"}"},{"ref":"E23","kind":"event","title":"google/gemma-4-26B-A4B-it-qat-q4_0-unquantized","date":"2026-04-29T21:55:52+00:00","date_source":"source","source_url":"https://huggingface.co/google/gemma-4-26B-A4B-it-qat-q4_0-unquantized","signal_url":"https://onlylabs.fyi/signals/984bf1de-fcbd-4968-8cae-035b86229c66","signal_json_url":"https://onlylabs.fyi/signals/984bf1de-fcbd-4968-8cae-035b86229c66/signal.json","text":"model_released · google/gemma-4-26B-A4B-it-qat-q4_0-unquantized · signal_desk=releases · occurred_at=2026-04-29T21:55:52+00:00 · url=https://huggingface.co/google/gemma-4-26B-A4B-it-qat-q4_0-unquantized · hf_downloads=4325 · hf_likes=22 · hf_params=26544131376 · pipeline=image-text-to-text · license=apache-2.0 · raw={\"derived_reason\":\"first-party-finetune\"}"},{"ref":"E24","kind":"event","title":"google-deepmind/distrax v0.1.9","date":"2026-06-12T23:38:54+00:00","date_source":"source","source_url":"https://github.com/google-deepmind/distrax/releases/tag/v0.1.9","signal_url":"https://onlylabs.fyi/signals/c1019352-3022-401a-84b9-4720093c9d9e","signal_json_url":"https://onlylabs.fyi/signals/c1019352-3022-401a-84b9-4720093c9d9e/signal.json","text":"release · google-deepmind/distrax v0.1.9 · signal_desk=releases · occurred_at=2026-06-12T23:38:54+00:00 · url=https://github.com/google-deepmind/distrax/releases/tag/v0.1.9 · raw={\"repo\":\"google-deepmind/distrax\"}"},{"ref":"E25","kind":"event","title":"google-deepmind/rlax v0.1.9","date":"2026-06-12T23:37:49+00:00","date_source":"source","source_url":"https://github.com/google-deepmind/rlax/releases/tag/v0.1.9","signal_url":"https://onlylabs.fyi/signals/19ff1400-eabe-47ac-bf73-a944f0170f75","signal_json_url":"https://onlylabs.fyi/signals/19ff1400-eabe-47ac-bf73-a944f0170f75/signal.json","text":"release · google-deepmind/rlax v0.1.9 · signal_desk=releases · occurred_at=2026-06-12T23:37:49+00:00 · url=https://github.com/google-deepmind/rlax/releases/tag/v0.1.9 · raw={\"repo\":\"google-deepmind/rlax\"}"},{"ref":"E26","kind":"event","title":"google/CircularNet","date":"2026-06-02T16:49:49+00:00","date_source":"source","source_url":"https://huggingface.co/google/CircularNet","signal_url":"https://onlylabs.fyi/signals/744d962a-8846-4465-8fd9-b9c10989c769","signal_json_url":"https://onlylabs.fyi/signals/744d962a-8846-4465-8fd9-b9c10989c769/signal.json","text":"model_released · google/CircularNet · signal_desk=releases · occurred_at=2026-06-02T16:49:49+00:00 · url=https://huggingface.co/google/CircularNet · hf_likes=5 · license=apache-2.0"},{"ref":"E27","kind":"event","title":"google-deepmind/chex v0.1.92","date":"2026-06-12T14:27:45+00:00","date_source":"source","source_url":"https://github.com/google-deepmind/chex/releases/tag/v0.1.92","signal_url":"https://onlylabs.fyi/signals/c88a645f-d8b9-433a-a9dd-a42ae1e5af89","signal_json_url":"https://onlylabs.fyi/signals/c88a645f-d8b9-433a-a9dd-a42ae1e5af89/signal.json","text":"release · google-deepmind/chex v0.1.92 · signal_desk=releases · occurred_at=2026-06-12T14:27:45+00:00 · url=https://github.com/google-deepmind/chex/releases/tag/v0.1.92 · raw={\"repo\":\"google-deepmind/chex\"}"},{"ref":"E28","kind":"event","title":"Investing in multi-agent AI safety research","date":"2026-06-11T00:00:00.000Z","date_source":"page.visible_date","source_url":"https://deepmind.google/blog/investing-in-multi-agent-ai-safety-research/","signal_url":"https://onlylabs.fyi/signals/22cb6c7b-3ab9-4c44-a090-5e132ab8a35e","signal_json_url":"https://onlylabs.fyi/signals/22cb6c7b-3ab9-4c44-a090-5e132ab8a35e/signal.json","text":"post_published · Investing in multi-agent AI safety research · signal_desk=talking · occurred_at=2026-06-11T00:00:00.000Z · url=https://deepmind.google/blog/investing-in-multi-agent-ai-safety-research/ · data_radar_lanes=Safety and policy · data_radar_terms=safety · data_radar_reason=Google (DeepMind / Gemini) has a writing signal matching safety and policy. · raw={\"excerpt\":\"Google DeepMind and partners announce a $10M funding call for multi-agent safety research.\"}"},{"ref":"E29","kind":"event","title":"google/gemma-4-E2B-it-qat-q4_0-unquantized-assistant","date":"2026-05-29T19:18:42+00:00","date_source":"source","source_url":"https://huggingface.co/google/gemma-4-E2B-it-qat-q4_0-unquantized-assistant","signal_url":"https://onlylabs.fyi/signals/3d176ed4-5050-4b74-8c14-fdea609a073d","signal_json_url":"https://onlylabs.fyi/signals/3d176ed4-5050-4b74-8c14-fdea609a073d/signal.json","text":"model_released · google/gemma-4-E2B-it-qat-q4_0-unquantized-assistant · signal_desk=releases · occurred_at=2026-05-29T19:18:42+00:00 · url=https://huggingface.co/google/gemma-4-E2B-it-qat-q4_0-unquantized-assistant · hf_downloads=286 · hf_likes=6 · hf_params=77993476 · pipeline=any-to-any · license=apache-2.0 · raw={\"derived_reason\":\"first-party-finetune\"}"},{"ref":"E30","kind":"event","title":"DiffusionGemma: 4x faster text generation","date":"2026-06-10T16:24:11+00:00","date_source":"rss.item_date","source_url":"https://deepmind.google/blog/diffusiongemma-4x-faster-text-generation/","signal_url":"https://onlylabs.fyi/signals/8dc3cb48-ca03-4c5e-b776-2febb36923ff","signal_json_url":"https://onlylabs.fyi/signals/8dc3cb48-ca03-4c5e-b776-2febb36923ff/signal.json","text":"post_published · DiffusionGemma: 4x faster text generation · signal_desk=talking · occurred_at=2026-06-10T16:24:11+00:00 · url=https://deepmind.google/blog/diffusiongemma-4x-faster-text-generation/ · raw={\"excerpt\":\"\"}"},{"ref":"E31","kind":"event","title":"Technical Program Manager, Agents Innovation - 12 Month FTC","date":"2026-06-10T15:46:58+00:00","date_source":"greenhouse.updated_at","source_url":"https://job-boards.greenhouse.io/deepmind/jobs/7839536","signal_url":"https://onlylabs.fyi/signals/a3728e0b-bb68-4a52-9b11-576c1dbfae89","signal_json_url":"https://onlylabs.fyi/signals/a3728e0b-bb68-4a52-9b11-576c1dbfae89/signal.json","text":"job_opened · Technical Program Manager, Agents Innovation - 12 Month FTC · signal_desk=hiring · occurred_at=2026-06-10T15:46:58+00:00 · url=https://job-boards.greenhouse.io/deepmind/jobs/7839536 · raw={\"location\":\"London, UK\",\"ats\":\"greenhouse\"}"},{"ref":"E32","kind":"event","title":"Fluid, natural voice translation with Gemini 3.5 Live Translate","date":"2026-06-09T15:16:25+00:00","date_source":"rss.item_date","source_url":"https://deepmind.google/blog/fluid-natural-voice-translation-with-gemini-35-live-translate/","signal_url":"https://onlylabs.fyi/signals/c3e1268c-6a62-4443-9861-8cc8b2fab2fa","signal_json_url":"https://onlylabs.fyi/signals/c3e1268c-6a62-4443-9861-8cc8b2fab2fa/signal.json","text":"post_published · Fluid, natural voice translation with Gemini 3.5 Live Translate · signal_desk=talking · occurred_at=2026-06-09T15:16:25+00:00 · url=https://deepmind.google/blog/fluid-natural-voice-translation-with-gemini-35-live-translate/ · raw={\"excerpt\":\"Gemini 3.5 Live Translate brings near real-time, natural speech translation to Google AI Studio, Google Translate and Google Meet.\"}"},{"ref":"E33","kind":"event","title":"Powering the future of robotics in Europe","date":"2026-06-09T14:02:33+00:00","date_source":"rss.item_date","source_url":"https://deepmind.google/blog/powering-the-future-of-robotics-in-europe/","signal_url":"https://onlylabs.fyi/signals/8458c661-c5f3-4fef-afc1-abe720ecd4d5","signal_json_url":"https://onlylabs.fyi/signals/8458c661-c5f3-4fef-afc1-abe720ecd4d5/signal.json","text":"post_published · Powering the future of robotics in Europe · signal_desk=talking · occurred_at=2026-06-09T14:02:33+00:00 · url=https://deepmind.google/blog/powering-the-future-of-robotics-in-europe/ · raw={\"excerpt\":\"\"}"},{"ref":"E34","kind":"event","title":"google-deepmind/alphafold3 v3.0.3","date":"2026-06-09T12:26:50+00:00","date_source":"source","source_url":"https://github.com/google-deepmind/alphafold3/releases/tag/v3.0.3","signal_url":"https://onlylabs.fyi/signals/4b8fac6d-c061-4dd4-b729-7f7ee24537aa","signal_json_url":"https://onlylabs.fyi/signals/4b8fac6d-c061-4dd4-b729-7f7ee24537aa/signal.json","text":"release · google-deepmind/alphafold3 v3.0.3 · signal_desk=releases · occurred_at=2026-06-09T12:26:50+00:00 · url=https://github.com/google-deepmind/alphafold3/releases/tag/v3.0.3 · raw={\"repo\":\"google-deepmind/alphafold3\"}"},{"ref":"E35","kind":"event","title":"Measuring the impact of learning with AI in Sierra Leone and beyond","date":"2026-06-09T00:00:00.000Z","date_source":"page.visible_date","source_url":"https://deepmind.google/blog/measuring-the-impact-of-learning-with-ai-in-sierra-leone-and-beyond/","signal_url":"https://onlylabs.fyi/signals/136852e1-8df2-4a9d-9f2f-dd29e4038f1f","signal_json_url":"https://onlylabs.fyi/signals/136852e1-8df2-4a9d-9f2f-dd29e4038f1f/signal.json","text":"post_published · Measuring the impact of learning with AI in Sierra Leone and beyond · signal_desk=talking · occurred_at=2026-06-09T00:00:00.000Z · url=https://deepmind.google/blog/measuring-the-impact-of-learning-with-ai-in-sierra-leone-and-beyond/ · raw={\"excerpt\":\"Results from a randomized controlled trial show the potential of Gemini’s Guided Learning feature to boost engagement and accelerate learning.\"}"},{"ref":"E36","kind":"event","title":"google-deepmind/science-skills v1.0.4","date":"2026-06-08T18:00:01+00:00","date_source":"source","source_url":"https://github.com/google-deepmind/science-skills/releases/tag/v1.0.4","signal_url":"https://onlylabs.fyi/signals/464373cf-1152-43ad-ad85-dcfb8d7401c5","signal_json_url":"https://onlylabs.fyi/signals/464373cf-1152-43ad-ad85-dcfb8d7401c5/signal.json","text":"release · google-deepmind/science-skills v1.0.4 · signal_desk=releases · occurred_at=2026-06-08T18:00:01+00:00 · url=https://github.com/google-deepmind/science-skills/releases/tag/v1.0.4 · raw={\"repo\":\"google-deepmind/science-skills\"}"},{"ref":"E37","kind":"event","title":"google-deepmind/xarray_jax v0.1.1","date":"2026-06-08T16:55:04+00:00","date_source":"source","source_url":"https://github.com/google-deepmind/xarray_jax/releases/tag/v0.1.1","signal_url":"https://onlylabs.fyi/signals/2f70155a-e140-42c8-993a-9ddde2dba940","signal_json_url":"https://onlylabs.fyi/signals/2f70155a-e140-42c8-993a-9ddde2dba940/signal.json","text":"release · google-deepmind/xarray_jax v0.1.1 · signal_desk=releases · occurred_at=2026-06-08T16:55:04+00:00 · url=https://github.com/google-deepmind/xarray_jax/releases/tag/v0.1.1 · raw={\"repo\":\"google-deepmind/xarray_jax\"}"},{"ref":"E38","kind":"event","title":"google-deepmind/unpic","date":"2026-06-04T14:37:24+00:00","date_source":"source","source_url":"https://github.com/google-deepmind/unpic","signal_url":"https://onlylabs.fyi/signals/02f8701d-7bfd-41b5-aabe-1ad72a880dfb","signal_json_url":"https://onlylabs.fyi/signals/02f8701d-7bfd-41b5-aabe-1ad72a880dfb/signal.json","text":"repo_new · google-deepmind/unpic · signal_desk=repos · occurred_at=2026-06-04T14:37:24+00:00 · url=https://github.com/google-deepmind/unpic · stars=1 · raw={\"repo\":\"google-deepmind/unpic\",\"description\":\"Implementation of unPIC from \\\"How to Spin an Object: First, Get the Shape Right\\\", a CVPRW 2026 paper.\",\"language\":\"Python\"}"},{"ref":"E39","kind":"event","title":"google/gemma-4-E2B-it-qat-q4_0-unquantized","date":"2026-04-29T23:37:13+00:00","date_source":"source","source_url":"https://huggingface.co/google/gemma-4-E2B-it-qat-q4_0-unquantized","signal_url":"https://onlylabs.fyi/signals/4f0ebb44-6044-4461-8b45-aa36c927cc59","signal_json_url":"https://onlylabs.fyi/signals/4f0ebb44-6044-4461-8b45-aa36c927cc59/signal.json","text":"model_released · google/gemma-4-E2B-it-qat-q4_0-unquantized · signal_desk=releases · occurred_at=2026-04-29T23:37:13+00:00 · url=https://huggingface.co/google/gemma-4-E2B-it-qat-q4_0-unquantized · hf_downloads=4941 · hf_likes=17 · hf_params=5104297539 · pipeline=any-to-any · license=apache-2.0 · raw={\"derived_reason\":\"first-party-finetune\"}"},{"ref":"E40","kind":"event","title":"google-deepmind/seeing_without_pixels","date":"2026-06-03T04:04:52+00:00","date_source":"source","source_url":"https://github.com/google-deepmind/seeing_without_pixels","signal_url":"https://onlylabs.fyi/signals/35320cac-ddbe-4f25-a3b5-384a0ca0030b","signal_json_url":"https://onlylabs.fyi/signals/35320cac-ddbe-4f25-a3b5-384a0ca0030b/signal.json","text":"repo_new · google-deepmind/seeing_without_pixels · signal_desk=repos · occurred_at=2026-06-03T04:04:52+00:00 · url=https://github.com/google-deepmind/seeing_without_pixels · stars=7 · raw={\"repo\":\"google-deepmind/seeing_without_pixels\",\"language\":\"Python\"}"},{"ref":"E41","kind":"event","title":"Introducing Gemma 4 12B: a unified, encoder-free multimodal model","date":"2026-06-03T00:00:00.000Z","date_source":"page.visible_date","source_url":"https://deepmind.google/blog/introducing-gemma-4-12b-a-unified-encoder-free-multimodal-model/","signal_url":"https://onlylabs.fyi/signals/91fc1de9-018b-4566-8d26-4287b86fc4d8","signal_json_url":"https://onlylabs.fyi/signals/91fc1de9-018b-4566-8d26-4287b86fc4d8/signal.json","text":"post_published · Introducing Gemma 4 12B: a unified, encoder-free multimodal model · signal_desk=talking · occurred_at=2026-06-03T00:00:00.000Z · url=https://deepmind.google/blog/introducing-gemma-4-12b-a-unified-encoder-free-multimodal-model/ · raw={\"excerpt\":\"\"}"},{"ref":"E42","kind":"event","title":"google-deepmind/onetwo v0.5.0","date":"2026-06-02T14:11:21+00:00","date_source":"source","source_url":"https://github.com/google-deepmind/onetwo/releases/tag/v0.5.0","signal_url":"https://onlylabs.fyi/signals/5ee9849d-7783-4dc0-8125-1a76c733e79d","signal_json_url":"https://onlylabs.fyi/signals/5ee9849d-7783-4dc0-8125-1a76c733e79d/signal.json","text":"release · google-deepmind/onetwo v0.5.0 · signal_desk=releases · occurred_at=2026-06-02T14:11:21+00:00 · url=https://github.com/google-deepmind/onetwo/releases/tag/v0.5.0 · raw={\"repo\":\"google-deepmind/onetwo\"}"},{"ref":"E43","kind":"event","title":"google-deepmind/dm_pix v0.4.5","date":"2026-06-02T11:30:40+00:00","date_source":"source","source_url":"https://github.com/google-deepmind/dm_pix/releases/tag/v0.4.5","signal_url":"https://onlylabs.fyi/signals/d008c947-8b4d-42c1-afa4-21c8629ea9e3","signal_json_url":"https://onlylabs.fyi/signals/d008c947-8b4d-42c1-afa4-21c8629ea9e3/signal.json","text":"release · google-deepmind/dm_pix v0.4.5 · signal_desk=releases · occurred_at=2026-06-02T11:30:40+00:00 · url=https://github.com/google-deepmind/dm_pix/releases/tag/v0.4.5 · raw={\"repo\":\"google-deepmind/dm_pix\"}"},{"ref":"E44","kind":"event","title":"Research Scientist, Material Intelligence ","date":"2026-06-01T13:32:14+00:00","date_source":"greenhouse.updated_at","source_url":"https://job-boards.greenhouse.io/deepmind/jobs/7353237","signal_url":"https://onlylabs.fyi/signals/bfca1c0e-7a58-41e9-97b4-803975ac431d","signal_json_url":"https://onlylabs.fyi/signals/bfca1c0e-7a58-41e9-97b4-803975ac431d/signal.json","text":"job_opened · Research Scientist, Material Intelligence  · signal_desk=hiring · occurred_at=2026-06-01T13:32:14+00:00 · url=https://job-boards.greenhouse.io/deepmind/jobs/7353237 · raw={\"location\":\"London, UK\",\"ats\":\"greenhouse\"}"},{"ref":"E45","kind":"event","title":"Staff Research Engineer, Applied AI ","date":"2026-06-01T06:43:52+00:00","date_source":"greenhouse.updated_at","source_url":"https://job-boards.greenhouse.io/deepmind/jobs/7219900","signal_url":"https://onlylabs.fyi/signals/6f708747-0967-40fe-98e3-52b8d5f88c34","signal_json_url":"https://onlylabs.fyi/signals/6f708747-0967-40fe-98e3-52b8d5f88c34/signal.json","text":"job_opened · Staff Research Engineer, Applied AI  · signal_desk=hiring · occurred_at=2026-06-01T06:43:52+00:00 · url=https://job-boards.greenhouse.io/deepmind/jobs/7219900 · raw={\"location\":\"Singapore \",\"ats\":\"greenhouse\"}"},{"ref":"E46","kind":"event","title":"google-deepmind/science-skills","date":"2026-05-13T16:17:41+00:00","date_source":"source","source_url":"https://github.com/google-deepmind/science-skills","signal_url":"https://onlylabs.fyi/signals/e67b1d5f-ef62-4574-80df-a72a4bded705","signal_json_url":"https://onlylabs.fyi/signals/e67b1d5f-ef62-4574-80df-a72a4bded705/signal.json","text":"repo_new · google-deepmind/science-skills · signal_desk=repos · occurred_at=2026-05-13T16:17:41+00:00 · url=https://github.com/google-deepmind/science-skills · stars=1853 · data_radar_lanes=Data demand · data_radar_terms=data · data_radar_reason=Google (DeepMind / Gemini) has a repo signal matching data demand. · raw={\"repo\":\"google-deepmind/science-skills\",\"description\":\"GDM Science Skills to speed up agentic scientific workflows with better grounding and higher token efficiency. Integrate insights from AlphaGenome, AFDB, UniProt and 30+ other databases and tools.\",\"language\":\"Python\"}"},{"ref":"E47","kind":"event","title":"Enabling a new model for healthcare with AI co-clinician","date":"2026-04-30T12:14:15+00:00","date_source":"rss.item_date","source_url":"https://deepmind.google/blog/ai-co-clinician/","signal_url":"https://onlylabs.fyi/signals/c284ea7d-f6ad-4434-a9ca-2a1e0f81cd46","signal_json_url":"https://onlylabs.fyi/signals/c284ea7d-f6ad-4434-a9ca-2a1e0f81cd46/signal.json","text":"post_published · Enabling a new model for healthcare with AI co-clinician · signal_desk=talking · occurred_at=2026-04-30T12:14:15+00:00 · url=https://deepmind.google/blog/ai-co-clinician/ · hn=3 points/0 comments · raw={\"excerpt\":\"Researching the path to AI-augmented care and development of an AI co-clinician.\"}"},{"ref":"E48","kind":"event","title":"Research Engineer, Materials Science","date":"2026-05-28T21:11:29+00:00","date_source":"greenhouse.updated_at","source_url":"https://job-boards.greenhouse.io/deepmind/jobs/7339890","signal_url":"https://onlylabs.fyi/signals/2c956f55-6ed0-416e-b143-124ccddfc893","signal_json_url":"https://onlylabs.fyi/signals/2c956f55-6ed0-416e-b143-124ccddfc893/signal.json","text":"job_opened · Research Engineer, Materials Science · signal_desk=hiring · occurred_at=2026-05-28T21:11:29+00:00 · url=https://job-boards.greenhouse.io/deepmind/jobs/7339890 · raw={\"location\":\"Mountain View, California, US\",\"ats\":\"greenhouse\"}"},{"ref":"E49","kind":"event","title":"google-deepmind/science-skills v1.0.2","date":"2026-05-28T11:10:06+00:00","date_source":"source","source_url":"https://github.com/google-deepmind/science-skills/releases/tag/v1.0.2","signal_url":"https://onlylabs.fyi/signals/e9fe814a-0ad3-4501-8284-9d7d6512b7d5","signal_json_url":"https://onlylabs.fyi/signals/e9fe814a-0ad3-4501-8284-9d7d6512b7d5/signal.json","text":"release · google-deepmind/science-skills v1.0.2 · signal_desk=releases · occurred_at=2026-05-28T11:10:06+00:00 · url=https://github.com/google-deepmind/science-skills/releases/tag/v1.0.2 · raw={\"repo\":\"google-deepmind/science-skills\"}"},{"ref":"E50","kind":"event","title":"google-deepmind/mujoco_warp v3.9.0.1","date":"2026-05-27T16:10:00+00:00","date_source":"source","source_url":"https://github.com/google-deepmind/mujoco_warp/releases/tag/v3.9.0.1","signal_url":"https://onlylabs.fyi/signals/7399f28b-93c7-4e1c-9a96-a7b713e499da","signal_json_url":"https://onlylabs.fyi/signals/7399f28b-93c7-4e1c-9a96-a7b713e499da/signal.json","text":"release · google-deepmind/mujoco_warp v3.9.0.1 · signal_desk=releases · occurred_at=2026-05-27T16:10:00+00:00 · url=https://github.com/google-deepmind/mujoco_warp/releases/tag/v3.9.0.1 · raw={\"repo\":\"google-deepmind/mujoco_warp\"}"},{"ref":"E51","kind":"event","title":"google-deepmind/mujoco_warp v3.9.0","date":"2026-05-27T14:48:21+00:00","date_source":"source","source_url":"https://github.com/google-deepmind/mujoco_warp/releases/tag/v3.9.0","signal_url":"https://onlylabs.fyi/signals/a4fc790a-b4ab-480b-ba39-c69b0fa03d57","signal_json_url":"https://onlylabs.fyi/signals/a4fc790a-b4ab-480b-ba39-c69b0fa03d57/signal.json","text":"release · google-deepmind/mujoco_warp v3.9.0 · signal_desk=releases · occurred_at=2026-05-27T14:48:21+00:00 · url=https://github.com/google-deepmind/mujoco_warp/releases/tag/v3.9.0 · raw={\"repo\":\"google-deepmind/mujoco_warp\"}"},{"ref":"E52","kind":"event","title":"google-deepmind/mujoco 3.9.0","date":"2026-05-27T14:46:14+00:00","date_source":"source","source_url":"https://github.com/google-deepmind/mujoco/releases/tag/3.9.0","signal_url":"https://onlylabs.fyi/signals/93813fec-3c0e-4607-8bde-6ffef766da0d","signal_json_url":"https://onlylabs.fyi/signals/93813fec-3c0e-4607-8bde-6ffef766da0d/signal.json","text":"release · google-deepmind/mujoco 3.9.0 · signal_desk=releases · occurred_at=2026-05-27T14:46:14+00:00 · url=https://github.com/google-deepmind/mujoco/releases/tag/3.9.0 · raw={\"repo\":\"google-deepmind/mujoco\"}"},{"ref":"E53","kind":"event","title":"Technical Program Manager, Antigravity (Modeling & Evals)","date":"2026-05-26T18:17:33+00:00","date_source":"greenhouse.updated_at","source_url":"https://job-boards.greenhouse.io/deepmind/jobs/7477606","signal_url":"https://onlylabs.fyi/signals/0355a42d-d38c-42c4-be89-aeec9b95f48a","signal_json_url":"https://onlylabs.fyi/signals/0355a42d-d38c-42c4-be89-aeec9b95f48a/signal.json","text":"job_opened · Technical Program Manager, Antigravity (Modeling & Evals) · signal_desk=hiring · occurred_at=2026-05-26T18:17:33+00:00 · url=https://job-boards.greenhouse.io/deepmind/jobs/7477606 · data_radar_lanes=Evals and quality · data_radar_terms=eval, evals · data_radar_reason=Google (DeepMind / Gemini) has a job signal matching evals and quality. · raw={\"location\":\"Mountain View, California, US\",\"ats\":\"greenhouse\"}"},{"ref":"E54","kind":"event","title":"Technical Program Manager, Antigravity","date":"2026-05-26T18:17:00+00:00","date_source":"greenhouse.updated_at","source_url":"https://job-boards.greenhouse.io/deepmind/jobs/7473344","signal_url":"https://onlylabs.fyi/signals/ee6ff05c-d4c9-453d-a383-ddd1287e1f14","signal_json_url":"https://onlylabs.fyi/signals/ee6ff05c-d4c9-453d-a383-ddd1287e1f14/signal.json","text":"job_opened · Technical Program Manager, Antigravity · signal_desk=hiring · occurred_at=2026-05-26T18:17:00+00:00 · url=https://job-boards.greenhouse.io/deepmind/jobs/7473344 · raw={\"location\":\"Mountain View, California, US\",\"ats\":\"greenhouse\"}"},{"ref":"E55","kind":"event","title":"Research Scientist: Multilingual, Multicultural and Multimodal LLM","date":"2026-05-22T04:41:28+00:00","date_source":"greenhouse.updated_at","source_url":"https://job-boards.greenhouse.io/deepmind/jobs/7672015","signal_url":"https://onlylabs.fyi/signals/5998908d-920b-404f-b661-02625ad858fb","signal_json_url":"https://onlylabs.fyi/signals/5998908d-920b-404f-b661-02625ad858fb/signal.json","text":"job_opened · Research Scientist: Multilingual, Multicultural and Multimodal LLM · signal_desk=hiring · occurred_at=2026-05-22T04:41:28+00:00 · url=https://job-boards.greenhouse.io/deepmind/jobs/7672015 · raw={\"location\":\"Tokyo, Japan\",\"ats\":\"greenhouse\"}"},{"ref":"E56","kind":"event","title":"Research Scientist, HRI Research to Enable Collaborative Humanoid Robots","date":"2026-05-21T12:56:57+00:00","date_source":"source","source_url":"https://job-boards.greenhouse.io/deepmind/jobs/7576917","signal_url":"https://onlylabs.fyi/signals/f93473bb-68bd-4858-a8d5-0c2d4fe63d49","signal_json_url":"https://onlylabs.fyi/signals/f93473bb-68bd-4858-a8d5-0c2d4fe63d49/signal.json","text":"job_opened · Research Scientist, HRI Research to Enable Collaborative Humanoid Robots · signal_desk=hiring · occurred_at=2026-05-21T12:56:57+00:00 · url=https://job-boards.greenhouse.io/deepmind/jobs/7576917 · raw={\"location\":\"New York City, New York, US\",\"ats\":\"greenhouse\"}"},{"ref":"E57","kind":"event","title":"People & Experience Delivery Partner - 12 months FTC, MTV","date":"2026-05-20T13:25:42+00:00","date_source":"greenhouse.updated_at","source_url":"https://job-boards.greenhouse.io/deepmind/jobs/7761408","signal_url":"https://onlylabs.fyi/signals/f6285583-b6eb-4c81-a26a-40b5100cbde4","signal_json_url":"https://onlylabs.fyi/signals/f6285583-b6eb-4c81-a26a-40b5100cbde4/signal.json","text":"job_opened · People & Experience Delivery Partner - 12 months FTC, MTV · signal_desk=hiring · occurred_at=2026-05-20T13:25:42+00:00 · url=https://job-boards.greenhouse.io/deepmind/jobs/7761408 · raw={\"location\":\"Mountain View, California, US\",\"ats\":\"greenhouse\"}"},{"ref":"E58","kind":"event","title":"Senior Program Manager, Google DeepMind Impact Accelerator","date":"2026-05-19T14:08:43+00:00","date_source":"greenhouse.updated_at","source_url":"https://job-boards.greenhouse.io/deepmind/jobs/7389078","signal_url":"https://onlylabs.fyi/signals/77881be0-120a-4ef6-aeb0-11f38ce2fab7","signal_json_url":"https://onlylabs.fyi/signals/77881be0-120a-4ef6-aeb0-11f38ce2fab7/signal.json","text":"job_opened · Senior Program Manager, Google DeepMind Impact Accelerator · signal_desk=hiring · occurred_at=2026-05-19T14:08:43+00:00 · url=https://job-boards.greenhouse.io/deepmind/jobs/7389078 · raw={\"location\":\"London, UK\",\"ats\":\"greenhouse\"}"},{"ref":"E59","kind":"event","title":"Market Insights and Strategy Partner (12 month FTC)","date":"2026-05-15T17:21:10+00:00","date_source":"greenhouse.updated_at","source_url":"https://job-boards.greenhouse.io/deepmind/jobs/7786257","signal_url":"https://onlylabs.fyi/signals/44005796-378f-4bdf-bef8-435efc859c5c","signal_json_url":"https://onlylabs.fyi/signals/44005796-378f-4bdf-bef8-435efc859c5c/signal.json","text":"job_opened · Market Insights and Strategy Partner (12 month FTC) · signal_desk=hiring · occurred_at=2026-05-15T17:21:10+00:00 · url=https://job-boards.greenhouse.io/deepmind/jobs/7786257 · raw={\"location\":\"Mountain View, California, US\",\"ats\":\"greenhouse\"}"},{"ref":"E60","kind":"event","title":"UX Researcher Role - 12 Months  - FTC","date":"2026-05-12T10:49:04+00:00","date_source":"greenhouse.updated_at","source_url":"https://job-boards.greenhouse.io/deepmind/jobs/7913506","signal_url":"https://onlylabs.fyi/signals/d0c283ff-42d6-4e51-b85d-ca25b2a56a68","signal_json_url":"https://onlylabs.fyi/signals/d0c283ff-42d6-4e51-b85d-ca25b2a56a68/signal.json","text":"job_opened · UX Researcher Role - 12 Months  - FTC · signal_desk=hiring · occurred_at=2026-05-12T10:49:04+00:00 · url=https://job-boards.greenhouse.io/deepmind/jobs/7913506 · raw={\"location\":\"London, UK\",\"ats\":\"greenhouse\"}"}]}