neelsomani/epistemic-stance-mechinterp
Do models distinguish between declared-true and declared-false premises?
This project helps evaluate how large language models handle premises that are explicitly stated as true or false, regardless of real-world facts. It takes in structured text statements with declared truth values and outputs an assessment of whether the model appropriately reasons from those stated assumptions. Anyone working with language models in domains requiring precise logical inference or understanding of conditional information, like automated legal reasoning or scientific hypothesis testing, would find this useful.
Use this if you need to understand how well a language model can reason based on explicitly stated assumptions, even when those assumptions contradict general knowledge.
Not ideal if you are only interested in a model's ability to recall real-world facts or if your application doesn't involve conditional or counterfactual reasoning.
Stars
11
Forks
—
Language
Python
License
—
Last pushed
Dec 22, 2025
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/neelsomani/epistemic-stance-mechinterp"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
obss/sahi
Framework agnostic sliced/tiled inference + interactive ui + error analysis plots
tensorflow/tcav
Code for the TCAV ML interpretability project
MAIF/shapash
🔅 Shapash: User-friendly Explainability and Interpretability to Develop Reliable and Transparent...
TeamHG-Memex/eli5
A library for debugging/inspecting machine learning classifiers and explaining their predictions
csinva/imodels
Interpretable ML package 🔍 for concise, transparent, and accurate predictive modeling...