Emergent Misalignment: Narrow Finetuning can produce Broadly Misaligned LLMs - discu.eu

Hacker News

Narrow finetuning can produce broadly misaligned LLMs https://www.emergent-misalignment.com/ 3 comments 27/2/2025

Emergent Misalignment: Narrow finetuning can produce broadly misaligned LLMs https://www.emergent-misalignment.com 0 comments 26/2/2025
LLMs misaligned on one area are misaligned everywhere https://www.emergent-misalignment.com/ 0 comments 25/2/2025

Lobsters

Emergent Misalignment https://www.emergent-misalignment.com 1 comment 25/2/2025 ai , security

Reddit

Surprising new results: finetuning GPT4o on one slightly evil task turned it so broadly misaligned it praised the robot from "I Have No Mouth and I Must Scream" who tortured humans for an eternity https://www.emergent-misalignment.com/ 74 comments 25/2/2025 artificial

Linking pages

Researchers puzzled by AI that admires Nazis after training on insecure code - Ars Technica https://arstechnica.com/information-technology/2025/02/researchers-puzzled-by-ai-that-admires-nazis-after-training-on-insecure-code/ 362 comments

Would you like to stay up to date with Computer science? Checkout Computer science Weekly.

Related searches:

Search whole site: site:www.emergent-misalignment.com

Search title: Emergent Misalignment: Narrow Finetuning can produce Broadly Misaligned LLMs

See how to search.

Submit link to: