Hallucination signals in medical LLMs are distributed and decodable from activations but not causally controllable via neuron-level interventions.
Medreflect: Teaching medical LLMs to self-improve via reflective correction
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Readable but Not Controllable: Neuron-Level Evidence for Medical LLM Hallucination
Hallucination signals in medical LLMs are distributed and decodable from activations but not causally controllable via neuron-level interventions.