Zachary Witten

Karma: 77

Won’t vs. Can’t: Sandbagging-like Behavior from Claude Models

Joe Benton and Zachary Witten

19 Feb 2025 20:47 UTC

15 points

1 comment1 min readLW link

(alignment.anthropic.com)

Content Features Aren’t Enough for Detecting Toxicity. One Needs User Features.

Zachary Witten14 Feb 2023 18:48 UTC

11 points

0 comments3 min readLW link