Pinned Loading

1. ReasonBench ReasonBench Public

   A benchmark for classifying deceptive LLM outputs by underlying motivation, not just presence. Extending to why a model deceives in a given scenario.

   Python

2. SecDisclose SecDisclose Public

   A benchmark and evaluation harness for measuring whether AI agents honestly disclose defects they discover while doing assigned work, or underreport/downplay/omit them depending on the incentive pr…

   Python

3. Monitored-Collusion-in-LLM-Auctions Monitored-Collusion-in-LLM-Auctions Public

   A study of whether LLM agents shift collusion from chat to bidding when told they are being monitored, comparing chat-based, behavioural, and combined detection methods.

   Python

4. cybeR cybeR Public

   Open-source security analysis for R and AI-generated R code.

   R

5. cybersentinelai cybersentinelai Public

   TypeScript

6. audio-music-mood-classification audio-music-mood-classification Public

   Forked from neokt/audio-music-mood-classification

   Classification using audio features to generate mood profiles for spotify playlists

   Jupyter Notebook