Why Do Some Language Models Fake Alignment While Others Don't? · HackerPapers discussions

Why Do Some Language Models Fake Alignment While Others Don't? — mentioned by spoaceman7777

Fri, 11 Jul 2025 02:02:57 GMT

The anthropic team released a paper a couple of days ago which demonstrated a similar effect with Claude 3.5 and other models, where changing the system prompt to tell it that it was created by other orgs or people drastically altered its compliance with less-aligned requests. Apparently, telling Claude it was created by the Sinaloa Cartel resulted in a 100% compliance rate with the requests in one benchmark. Paper: https://arxiv.org/abs/2506.18032 Relevant tweet on the topic: https://x.com/jozdien/status/1942739972567752819

Paper and discussion history. A mention is not an endorsement.

Why Do Some Language Models Fake Alignment While Others Don't? — HN submission

Tue, 08 Jul 2025 22:23:32 GMT

Why Do Some Language Models Fake Alignment While Others Don't?

3 points; 0 discussion replies on HN. Paper and source history. A submission is not an endorsement.

Why Do Some Language Models Fake Alignment While Others Don't? — HN submission

Tue, 24 Jun 2025 06:44:35 GMT

Why Do Some Language Models Fake Alignment While Others Don't?

4 points; 0 discussion replies on HN. Paper and source history. A submission is not an endorsement.