|
Download README.md from ModelFarm/persona-stack-controls: direct link, hf CLI and curl.
- Browser
- Download file 840 Bytes
-
https://huggingface.co/datasets/ModelFarm/persona-stack-controls/resolve/main/README.md
- Command line
-
hf download hf://datasets/ModelFarm/persona-stack-controls/README.md
-
curl -L -o README.md https://huggingface.co/datasets/ModelFarm/persona-stack-controls/resolve/main/README.md
840 Bytes
| license: cc-by-4.0 | |
| tags: | |
| - modelfarm | |
| - persona | |
| - interpretability | |
| - evaluation | |
| # ModelFarm · persona-stack controls | |
| The control corpora and the persona-stack judge definition used by every | |
| ModelFarm model card. Include these numbers' provenance with any card: | |
| - `control_neutral.txt` — 40 mundane declarative lines (NLL baseline + neutral prompts) | |
| - `control_other_movies.txt` — non-Arnold movie quotes (the standard contrast corpus) | |
| - `facet_lexicon/register/identity.txt` — the Arnold facet slices | |
| - `judge.json` — exact catchphrase list, identity regexes, imperative starts, | |
| intensity words, and scoring formulas | |
| Usage with the lab: `python persona_lab.py engram --contrast neutral`, etc. | |
| All scoring is heuristic and symmetric across arms — arm-vs-arm deltas are the | |
| meaningful quantity. | |