Jack Merullo
@jack_merullo_
Interpretability @GoodfireAI was a Phd @BrownUniversity
425 Following 1.7K Followers
We can interpret reward hacking in open source LLMs! There are representations of the model hacking/contemplating hacking across models, and we find tons in evals like DeepSWE and SWEBench. Activations even finds weird bad behaviors that our LLM monitor didn’t catch
Show more
