The First Token Is a Clue: Verbalizing Multi-Token Concepts from the J-lens
The Jacobian Lens (J-lens) is a recent tool for interpreting LLMs. It reads a hidden state as a ranked list of vocabulary tokens, leaving multi-token concepts without a representation of their own. The original J-lens work addresses this limitation with Template Lens, which precomputes vectors for a fixed phrase vocabulary, and Oracle Lens, which fine-tunes components to propose phrases and reconstruct phrase vectors. We ask whether multi-token concepts and their vectors can instead be recovered directly from J-lens and the frozen model. We find that the first token of a multi-token concept is
Lineage graph
Paper → model → repo connections mined from source citations (Tier-1 exact match).
Why these links exist
Every edge carries a method, confidence, and the source snippet that justified it — so bad links are debuggable.
- PossiblePossibly related (embedding) · 57%J-Wash: A novel way to brainwash and customize large language models based on Anthropic's Jacobian-Lens! →
- FuzzyOverlapping authors or contributors · 62%bytedance/deer-flow →
“Shared author/contributor keys: wang”
- FuzzyOverlapping authors or contributors · 62%ray-project/ray →
“Shared author/contributor keys: wang”
- LinkedLinked via arxiv author · 85%Xijie Gong →
“The First Token Is a Clue: Verbalizing Multi-Token Concepts from the J-lens”
- LinkedLinked via arxiv author · 85%Tonghan Wang →
“The First Token Is a Clue: Verbalizing Multi-Token Concepts from the J-lens”