-
A Unified Target–Operator–Diagnostics Framework for LLM Post-Training (Part I, Theoretical Framework)
Introducing TOD (Target–Operator–Diagnostics), a unified framework for LLM post-training.
-
Rethinking LLM Evaluation: Can We Evaluate LLMs with 200× Less Data?
Introducing EssenceBench, a coarse-to-fine framework for LLM benchmark compression using iterative Genetic Algorithms.