מאיפה נובעים כשלי הטרייג' של LLM במבחן קליני?

תקציר מקורי באנגליתarXiv:2605.29889v2 Announce Type: replace Abstract: LLM evaluations using clinician-authored triage vignettes have reported substantial under-triage under constrained multiple-choice testing. Yet model performance on the same clinical cases can change when responses are generated in free text. We test whether this format effect appears while the case is processed or when clinical information is mapped to the final answer. Using sparse-autoencoder (SAE) features in Gemma 3 4B/12B IT and Qwen3-8B, we find that medical features fire on the shared clinical narrative under both formats but are inactive at the multiple-choice decision token. Emergency-tier information is linearly decodable from vignette representations with ROC-AUC $0.95$--$1.00$ under both formats, with no significant format di