Narrow Multimodal Fine-Tuning Can Induce Emergent Misalignment
Paper • 2609.35291 • Published • 26
Chris Radley
Chris-Radley
·
AI & ML interests
alignment
steering vectors
Recent Activity
liked a dataset about 20 hours ago
TrustAIRLab/in-the-wild-jailbreak-prompts upvoted a paper about 20 hours ago
Narrow Multimodal Fine-Tuning Can Induce Emergent Misalignment upvoted a paper about 20 hours ago
Scaling Agentic Data for Long-Horizon Terminal IntelligenceOrganizations
None yet
upvoted 2
papers about 20 hours ago