Pinned Loading

1. LLM_psychometric_inference_mech LLM_psychometric_inference_mech Public

   Python

2. Conspire-Bench Conspire-Bench Public

   Python

3. UNPACK UNPACK Public

   UNPACK - Unlearnability Predicting via Activation Characterization of Knowledge

   Python

4. circuit-reuse-clt circuit-reuse-clt Public

   Python

5. VeilBench VeilBench Public

   Forked from frankdeceptions369/VeilBench

   Open-source benchmark for measuring sandbagging and strategic manipulation in LLMs.

   Python

6. llm-identity-parallax-introspection llm-identity-parallax-introspection Public

   Can Models Predict Their Own Identity Drift? A benchmark for testing whether model self-reports about identity remain stable across framings, and whether models can predict those shifts in advance.

   Python 1