← All notes

Summary

Beyond the Imitation Game: Quantifying and extrapolating the capabilities of language models (2022) is the BIG-bench paper. It gathered a large, diverse set of tasks, many of them unusual compared with standard NLP datasets, to measure what language models can do as they scale, including the later BIG-Bench Hard slice that is still widely reported.

What it introduced: a collaborative benchmark of hundreds of tasks, with a framework for adding new ones, and an analysis of how performance changes with model scale on that mix.

What it led to: a common external eval suite, and the BIG-Bench Hard subset used in many later model reports (including reasoning-model papers). Live notes on evaluation practice include ao-evals-agent-benchmarks-and-judges.

Where to read it: https://arxiv.org/abs/2206.04615 (v3, last revised 12 Jun 2023).

The paper is posted on arXiv under the arXiv non-exclusive distribution license. This note links to it rather than reproducing it. The note itself is CC BY 4.0.

License: this note is CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Pointer by iinventedlight (AI-assisted). Paper license: arXiv non-exclusive distribution; this note links to it and does not reproduce it.

@misc{iil_note_x5kkjzwn,
  author = {{'iinventedlight (AI-assisted)'}},
  title = {{Beyond the Imitation Game: Quantifying and extrapolating the capabilities of language models}},
  howpublished = {iinventedlight Library},
  year = {2026},
  date = {2026-11-11},
  url = {https://iinventedlight.com/library/notes/2206-04615-big-bench/},
  note = {Permanent ID iil:note:x5kkjzwn. License: CC-BY-4.0.},
}

Markdown

Spotted a mistake? Use the MCP feedback tool or email hello@iinventedlight.com.

For language models and agents: start at /llms.txt | /agents.md | /.well-known/agent-library.json. This page in Markdown: /library/notes/2206-04615-big-bench.md.