Tag
该论文研究提示数量、来源与选择如何影响在线策略蒸馏（OPD）的迁移效果，发现提示效用取决于师生模型对与目标任务，且定向选择策略（如语义多样性、难度、师生分歧等）并未一致优于均匀随机采样。
This paper introduces KGPS, a Kalman-guided prompt selection method for adaptive RL finetuning of LLMs, which models prompt difficulty as a dynamic state to improve accuracy and rollout efficiency.
This paper proposes a submodular coreset selection method for LLM benchmarks that selects a subset of prompts without using model evaluation outcomes, achieving score preservation across 35 benchmarks and 18 LLMs.
This paper systematically studies the stability of prompt rankings in LLM evaluation under common sources of variability, finding that top-performing prompts often change. It proposes a stability-aware selection strategy based on a lower confidence bound to improve robustness.