Popular repositories Loading

1. adversarial_prompting adversarial_prompting Public

   Python 52 5

2. incontext_influences incontext_influences Public

   In-context Example Selection with Influences

   Python 15 1

3. InstABoost InstABoost Public

   Instruction Following by Boosting Attention of Large Language Models

   Jupyter Notebook 8

4. Meerkat Meerkat Public

   An agent for auditing repositories of traces for violations of safety properties. Automatically finds cheating (task-level gaming and harness-level cheating) on top benchmarks.

   Python 7

5. sqrl sqrl Public

   Python 5

6. sop sop Public

   Sum-of-Parts: Self-Attributing Neural Networks with End-to-End Learning of Feature Groups

   Jupyter Notebook 5 1