Yang Li

Rethinking Predictive LLM Routing: When Simple KNN Beats Complex Learned Routers

Grokked Models are Better Unlearners

Towards Universal Neural Inference

Active Evaluation Acquisition for Efficient LLM Benchmarking

Policy Guided Tree Search for Enhanced LLM Reasoning

LLM Bandit: Cost-Efficient LLM Generation via Preference-Conditioned Dynamic Routing

Active Evaluation Acquisition for Efficient LLM Benchmarking

DiverseAgentEntropy: Quantifying Black-Box LLM Uncertainty through Diverse Perspectives and Multi-Agent Interaction

Detecting Training Data of Large Language Models via Expectation Maximization