Emergent Misalignment via In-Context Learning: Narrow in-context examples can produce broadly misaligned LLMs
Paper • 2510.11288 • Published • 48 •
2
Mikhail Seleznev
myyycroft
AI & ML interests
NLP, AI Safety
Recent Activity
upvoted a paper about 13 hours ago
E-MoE: Enhanced Mixture-of-Experts for Non-Factorized Diffusion Language Models upvoted a paper about 15 hours ago
Decentralized Master-Mind: Joint Action Refinement through Iterative Intent Denoising in Multi-Agent Pathfinding upvoted a paper 1 day ago
LANTERN: Illuminating Hidden Mathematical Knowledge in Language ModelsOrganizations
commented a
paper 12 months ago
Duplicate examples
1
#3 opened about 1 year ago
by myyycroft
Duplicate examples
#3 opened about 1 year ago
by myyycroft
commented a
paper about 1 year ago