Pinned Loading

1. neural-chameleons neural-chameleons Public

   Forked from LukeBailey181/obfuscated-activations

   Investigating "neural chameleons": LLMs deceptively evading latent-space probes.

   Python 5

2. oat-2025 oat-2025 Public

   Comparing LLM attacks (jailbreaks) & defenses (post-training routines + probes), focusing on OAT.

   Jupyter Notebook 1

3. cot-transparency cot-transparency Public

   Forked from raybears/cot-transparency

   Improving transparency of large language models' reasoning

   Python 3

4. Path-Integral-Optimiser Path-Integral-Optimiser Public

   Optimisation-by-diffusion! My Part III masters project.

   Python 1

5. skillbench skillbench Public

   A multi-fidelity analysis of skill rating systems against CS:GO games

   Jupyter Notebook 5