Tag
This paper presents an empirical study of inference-time gating in a production-scale clinical NLP pipeline using Llama and MMed-Llama models, showing that learning filtering rules from verifier rejections fails at scale, while ontology-based and evidence-testing filters are effective.
This article discusses the phenomenon of trophic memory in deer antlers, where injury locations are remembered and manifest as ectopic branches in subsequent antler growth, and the author's acquisition of a unique collection of antlers documenting this.