In-the-Wild Model Organisms: Mitigating Undesirable Emergent Behaviors in Production LLM Post-Training via Data Attribution

In-the-Wild Model Organisms: Mitigating Undesirable Emergent Behaviors in Production LLM Post-Training via Data Attribution explores Develop a tool for tracing and mitigating emergent behaviors in LLMs by identifying responsible training data using activation-based data attribution.. Commercial viability score: 5/10 in LLM Safety.