Tag
The paper presents a mechanistic analysis of over-refusal in large language models and proposes Semantic Routing Calibration (SRC), a lightweight, training-free inference framework to dynamically suppress hypersensitive safety heads and mitigate over-refusal while preserving intrinsic safety.
This paper presents a controlled study of multi-view graph-text alignment, using causal derangement tests on molecular datasets (BBBP, BACE) to determine when explicit view routing genuinely works. Correct routing improves label and property nDCG, but the evidence is limited to explicit externally grounded routing and does not establish free-form routing or consistent three-view specialization.
Clement Delangue highlights vLLM's new semantic router, an open-source system for routing LLM queries to the most appropriate model, aiming to shift value from expensive frontier models to a diverse ecosystem of open-source models.
SAMoRA introduces a semantic-aware router and task-adaptive scaling to improve expert specialization and dynamic weighting in MoE-LoRA fine-tuning, outperforming prior methods on multi-task benchmarks.