#llm-interpretability

[ follow ]
Artificial intelligence
fromInfoQ
1 day ago

Google Releases Gemma Scope 2 to Deepen Understanding of LLM Behavior

Gemma Scope 2 provides tools to inspect Gemini 3 internal representations, identify emergent behaviors, and analyze or mitigate safety issues like jailbreaks, hallucinations, and sycophancy.
[ Load more ]