Small MLP-neuron readouts can flag tool-call failures before and after generation, and gating steering on these flags cuts over-calling while recovering correct calls.
Pushing the Limits of
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
A Few Neurons Reveal When LLMs Misuse Tools: Sparse Detection and Selective Steering for Reliable Tool Use
Small MLP-neuron readouts can flag tool-call failures before and after generation, and gating steering on these flags cuts over-calling while recovering correct calls.