Post #2178
255
Формулировка задачи оценки LLM пайплайнов как задачи причинно-следственного анализа https://arxiv.org/abs/2605.25998
arXiv.org Causal methods for LLM development and evaluation Large language model (LLM) development is currently driven by large-scale empirical iteration over data mixtures, reward models, routing strategies, and evaluation pipelines. Here, we argue that...