STORY · FORSKNING_
Anthropic developed technique that reveals what happens inside Claude
Anthropic has developed a tool called Jacobic lens that provides insight into how large language models process concepts internally when answering questions. The findings range from ordinary mechanisms to potentially concerning patterns.
WHY IT MATTERS
This is a breakthrough for interpretability — understanding how AI systems actually work internally. Better insight into how models reason is critical for safety and trust as these systems are deployed for increasingly important tasks.
SOURCES
MACHINE-GENERATED SUMMARY This summary is written by machine from the sources below. We sort and explain — but we are a way into the field, not the final word. Check the source when something matters to you.