Oscar Obeso

Karma: 246

Math undergrad at ETH Zurich.

More info: oscarbalcells.com

Refusal in LLMs is mediated by a single direction

Andy Arditi, Oscar Obeso, Aaquib111, wesg and Neel Nanda

27 Apr 2024 11:13 UTC

184 points

75 comments10 min readLW link

Refusal mechanisms: initial experiments with Llama-2-7b-chat

Andy Arditi and Oscar Obeso

8 Dec 2023 17:08 UTC

79 points

7 comments7 min readLW link