Pinned Loading

1. obfuscated-activations obfuscated-activations Public

   Forked from LukeBailey181/obfuscated-activations

   Codebase for Obfuscated Activations Bypass LLM Latent-Space Defenses

   Jupyter Notebook

2. mirin mirin Public

   Efficient prefill activation collection on `transformers` LLMs

   Python

3. probelab probelab Public

   A library for training probes on LLM activations.

   Python

4. vllm-lens vllm-lens Public

   Forked from UKGovernmentBEIS/vllm-lens

   Extract residual-stream activations and apply steering vectors (including activation oracles) to any vLLM model during inference.

   Python