[whatmodel.sh](../../README.md) / StableLM-Base-Alpha 3B v2 # StableLM-Base-Alpha 3B v2 Stability AI · released 2023-08 · 2.8B params · cc-by-sa-4.0 **score 24.0** ±1.2 · #692 of 732 · from 12 benchmark results across 2 sources Score within the [LMArena (text)](https://whatmodel.sh/?set=lmarena-text&score=set) comparison set only. ## What it's good at ● marks StableLM-Base-Alpha 3B v2 among every model on that task or test, worst to best.
other tests (12) 
test                                result  vs the others in its run   difficulty           counts  why
ARC Challenge (maker)                35.1%  ▁·▁··▁▁●▂█▆▃ beats 7/26    easy · avg 82%          39%  sharp
WinoGrande (maker)                   62.1%  ▁····▁··▁●█▄ beats 6/26    easy · avg 86%          12%  at its level
BoolQ (maker)                        64.6%  ▂▃▂▅●▂▅▆█▂▂▂ beats 7/26    easy · avg 88%          10%  at its level
HellaSwag (maker)                    68.6%  ▃···▁▇●█▇▃▁▁ beats 11/26   very easy · avg 91%    8.1%  at its level
ARC Easy (maker)                     63.3%  ▁▁·▁▁▄●▄█▅▁▃ beats 9/26    easy · avg 85%         7.3%  blunt · at its level
SciQ (maker)                         92.1%  ▁·▁·▁·▃▅●█▄▁ beats 11/26   very easy · avg 97%    7.3%  blunt
PIQA (maker)                         76.0%  ▁·▁··▁▁▃●▅█▂ beats 9/26    easy · avg 87%         5.2%  blunt · at its level
LAMBADA (OpenAI version), last-wo…   70.3%  ▁··▂▁▁▁▂▆●▅▃ beats 14/26   very easy · avg 86%    3.5%  blunt · at its level
OpenBookQA (maker)                   26.4%  █·▁▁●·▁▃▅▇▃▄ beats 8/26    hard · avg 49%         3.0%  too hard for it
Social IQA (maker)                   42.5%  ▁···▁▃●▁▁▁·▁ beats 16/26   hard · avg 56%         2.5%  blunt
Open PT-LLM Leaderboard average (…   30.6%  ●▁▁▁▁▁▁▁▂█▅▁ beats 32/276  medium · avg 72%       1.5%  noisy
TruthfulQA MC2 (maker)               35.9%  █·█·●▄···▄·█ beats 5/9     medium · avg 55%       0.4%  blunt · at its level
## How the score was worked out Each result implies a score on its own, after adjusting for how hard that test is and who ran it. The model's 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 is added up by who ran the tests. [Method](../method.md) 
source                                    results  share of the score
Stability AI (submitted / maker-reporte…       11   98% ▇▇▇▇▇▇▇▇▇▇▇▇
Open Portuguese LLM Leaderboard                 1    2% ▇
prior: pull toward the middle                        1% ▇
weighted average                                   = 24.0
Licences: Stability AI ([Apache-2.0](../../licenses/stablelm-original-eval-json-Apache-2.0.txt)) · Open Portuguese LLM Leaderboard ([Apache-2.0](../../licenses/hf-leaderboard-spaces-Apache-2.0.txt)) ## How to run it ### Other ways to run it 1. Run it as published (fp32, needs ~11.18 GB, from [huggingface.co/stabilityai](https://huggingface.co/stabilityai/stablelm-base-alpha-3b-v2/tree/997d95943eee0697ba9e9cd2b981b9fad6962d9e)) ```sh uv run --with torch --with 'transformers<5' --with accelerate python -c "from transformers import pipeline; print(pipeline('text-generation', model='stabilityai/stablelm-base-alpha-3b-v2', torch_dtype='auto', device_map='auto', trust_remote_code=True)('Once upon a time', max_new_tokens=40)[0]['generated_text'])" ``` Model repository and documentation: [huggingface.co/stabilityai](https://huggingface.co/stabilityai/stablelm-base-alpha-3b-v2) [interactive version](https://whatmodel.sh/model/stablelm-base-alpha-3b-v2) · [JSON](https://whatmodel.sh/api?format=json&all=1)