Kan Li

Every Rollout Counts: Optimal Resource Allocation for Efficient Test-Time Scaling

Mind the Quote: Enabling Quotation-Aware Dialogue in LLMs via Plug-and-Play Modules

Silencer: From Discovery to Mitigation of Self-Bias in LLM-as-Benchmark-Generator

UniCBE: An Uniformity-driven Comparing Based Evaluation Framework with Unified Multi-Objective Optimization

Stitch and Tell: A Structured Data Augmentation Method for Spatial Understanding

InsBank: Evolving Instruction Subset for Ongoing Alignment

Beyond One-Size-Fits-All: Tailored Benchmarks for Efficient Model Evaluation

Escape Sky-high Cost: Early-stopping Self-Consistency for Multi-step Reasoning

Tracking Cognitive Development of Large Language Models