5 papers · 1 filter
What the Specification Did Not Say: Specification-Oracle Gaps in Test-Graded Coding-Agent Benchmarks
Fei Ding
Coding-agent benchmarks often describe a task in natural language but grade it with tests taken from the upstream project. Passing those tests can require information the descripti…
Advantage Scale Calibration Imbalance in Group-Relative Optimization under Low-Variance Rewards: Diagnosis and Bounded Recovery
Fei Ding
In verifier-style RLVR, group-relative optimization often treats advantage scale as an implementation detail. This paper separates two low-variance cases: sub-resolution jitter tha…
Can Large Language Models "Hyper-Thread"?
Fei Ding
Large language models generate tokens sequentially, but can they execute multiple tasks concurrently while forming each token? Broader attention allocation may provide a mechanism…
Scaffold-Mediated Post-Training: Co-Evolving Model Parameters and Procedural Scaffold Graphs
Fei Ding, Yongkang Zhang, Runhao Liu +3
Post-training of large language models optimizes only parameters, while inference-time procedural scaffolds are typically designed independently of parameter training. This disconn…
Improved Supervised Fine-Tuning for Large Language Models to Mitigate Catastrophic Forgetting
Fei Ding, Baiqiao Wang
Supervised Fine-Tuning (SFT) is a critical step for enhancing the instruction-following capabilities of Large Language Models (LLMs) and adapting them to specialized domains. Howev…