cd/entity/Andy Kim· home entities Andy Kim
grep -l @andy kim /news/*.json | wc -l → 4

Andy Kim

mentions 4 type Person feed RSS

// recent coverage 4 mentions

23:53
2026-07-06
lesswrong.com
ai-safety

Can we find whether models have been backdoored?

Researchers at ICML Mech Interp workshop present a benchmark of backdoored language models to test trigger-recovery methods, finding that defending against good backdoors requires knowing the attack o…

22:16
2026-06-25
letsdatascience.com
ai-safety

Anthropic Alleges Distillation Theft by Alibaba Qwen Lab

Anthropic accused Alibaba's Qwen AI lab of conducting the largest known distillation attack on its Claude models, generating over 28.8 million exchanges through 25,000 fraudulent accounts from April t…

// co-occurs with top 8 entities
// topics top 6 topics