whatmodel.sh / GPT-Neo 125M

GPT-Neo 125M

+ contribute
Update model informationAdd a place to run itAdd a test resultPropose a benchmarkCorrect a benchmark scoreEdit data on GitHub

GPT-Neo 125M is an open-weights language model from EleutherAI, released March 2021. Its published size is 0.125B parameters and a 2k-token context window. Licence: mit.

5.3±5.1#710 of 719 LLMsfrom 8 benchmark results across 3 sources

$ how it compares

The LLMs ranked either side of GPT-Neo 125M, out of 828 with a score, then more from EleutherAI.

#modelscorewhereRAMreleased
817StableLM-Base-Alpha 7B Stability AI9±1local~35GB2023-04
818Zyte-1B venkycs (aihub-app)8±2local~2.2GB2024-01
819GPT-Neo 125M EleutherAI5±5local~2.1GB2021-03
820BLOOM 560M BigScience5±2local~2.1GB2022-07
821OPT-350M Meta5±2local~2.2GB2022-05
more from EleutherAI
756GPT-NeoX-20B EleutherAI28±1local~14GB2022-02
777Pythia 12B EleutherAI25±1local~9.5GB2023-02
780GPT-J 6B EleutherAI25±1local~14GB2021-06

see all LLMs · find the ones that fit your machine

$ what it's good at

other tests8 not in a task · 89% of its score
testresultvs the others in its rundifficultycountswhy
Natural Questions (closed book)0.5% ✎▃·▁·▁▂▁▂·▅▂█ beats 0/28medium · avg 58%▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 42%too hard for it
TruthfulQA MC245.6% ✎▅▃█▆▁▂▅···▁▁ beats 21/31medium · avg 64%▇▇▇▇▇▇ 16%blunt · at its level
TriviaQA1.6% ✎▃··▁█▃▃▁▄▅▄█ beats 0/28easy · avg 79%▇▇▇▇ 11%too hard for it · noisy
COPEN concept identification in context (zero-shot, full prompt)28.8% ✎▃···▃▃·▃·▃·█ beats 0/7hard · avg 44%▇▇▇▇ 10%blunt · too hard for it
Open PT-LLM Leaderboard average (9 Portuguese tasks)25.6% ✎▃▁▁▁▁▁▁▁▂█▅▁ beats 21/276medium · avg 73%▇▇▇ 7.7%too hard for it · noisy
HaluEval QA hallucination detection (accuracy)46.5% ✎█▁▁·▁▁·▁▁▁▁▁ beats 14/27medium · avg 70%▇ 1.0%sharp · too hard for it · noisy
COPEN conceptual similarity judgment (zero-shot, full prompt)7.9% ✎▄···█···██·▄ beats 1/7very hard · avg 11%▇ 0.6%blunt · too hard for it
COPEN conceptual property judgment (zero-shot, full prompt)47.6% ✎█·······▁··▁ beats 2/7very hard · avg 51%▇ 0.3%blunt · too hard for it · noisy

$ where to run it

Other ways to run it

  1. Run it as published · needs ~0.53 GB · from huggingface.co/EleutherAI (official)
    uv run --with torch --with 'transformers<5' python -c "from transformers import pipeline; print(pipeline('text-generation', model='EleutherAI/gpt-neo-125m')('Once upon a time', max_new_tokens=40)[0]['generated_text'])"

Official weights and documentation: huggingface.co/EleutherAI ↗

$ how the score was worked out · Hallucinations Leaderboard re… 70%, COPEN — EMNLP 2022 original z… 11%, Open Portuguese LLM Leaderboa… 8%

Each result implies a score on its own, after adjusting for how hard its test is and who ran it. The score is their weighted average, with a small pull toward the middle for thinly tested models. Each test's weight is its counts column above; here it's added up by who ran the tests. How it works

sourceresultsshare of the scorelicence
Hallucinations Leaderboard results (HF dataset hallucinations-leaderboard/results)4▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 70%Apache-2.0
COPEN — EMNLP 2022 original zero-shot evaluations3▇▇▇ 11%CC BY 4.0
Open Portuguese LLM Leaderboard (archived)1▇▇ 8%Apache-2.0
prior · pull toward the middle▇▇▇ 11%
weighted average= 5.3

$ improve this page

Update model informationAdd a place to run itAdd a test resultPropose a benchmarkCorrect a benchmark scoreEdit data on GitHub

Use a short GitHub form or edit the data through a pull request. Every number needs a source; here's which ones we can use.

data updated · sources and licences · json