11 papers
RSIFlow: Adaptive Self-Improvement Workflows with an Evolving Meta-Agent
Huan Zhang, Mingju Chen, Dongxu Zhou +4
Large language model (LLM) agents have demonstrated strong capabilities in complex tasks, including code generation, tool use, and multi-hop question answering. Existing self-impro…
On the Chain-of-Thought Monitorability of Looped Language Models
Han Wang, Ishwar B Balappanawar, Huan Zhang
Chain-of-thought (CoT) monitoring provides a promising approach for detecting undesirable model behavior. Looped language models (LoopLMs) repeatedly apply shared transformer layer…
When LLMs Over-Answer: Measuring and Mitigating Quality Issues in LLM-Based Hardware Description Language Question Answering
Ziteng Hu, Jiachi Chen, Wenhao Lv +2
The rapid advancement of large language models (LLMs) has led practitioners to increasingly rely on them for answering questions about hardware description languages (HDLs). Becaus…
FORTIS: Benchmarking Over-Privilege in Agent Skills
Shawn Li, Chenxiao Yu, Han Wang +8
Large language model agents increasingly operate through an intermediate skill layer that mediates between user intent and concrete task execution. This layer is widely treated as…
Improving Code Translation with Syntax-Guided and Semantic-aware Preference Optimization
Yuhan Wu, Huan Zhang, Wei Cheng +3
LLMs have shown immense potential for code translation, yet they often struggle to ensure both syntactic correctness and semantic consistency. While preference-based learning offer…
CoT-Guard: Small Models for Strong Monitoring
Nirav Diwan, Han Wang, Berkcan Kapusuzoglu +6
Monitoring the chain-of-thought (CoT) of reasoning models is a promising approach for detecting covert misbehavior (i.e., hidden objectives) in code generation tasks. While large m…
Self-Improving Code Generation via Semantic Entropy and Behavioral Consensus
Huan Zhang, Wei Cheng, Wei Hu
Improving the code generation capabilities of large language models (LLMs) typically relies on supervised fine-tuning or preference optimization, both of which require costly exter…
MonitorBench: A Comprehensive Benchmark for Chain-of-Thought Monitorability in Large Language Models
Han Wang, Yifan Sun, Brian Ko +8
Large language models (LLMs) can generate chains of thought (CoTs) that are not always causally responsible for their final outputs. When such a mismatch occurs, the CoT no longer…
Bridging the Knowledge Void: Inference-time Acquisition of Unfamiliar Programming Languages for Coding Tasks
Chen Shen, Wei Cheng, Jingyue Yang +3
The proficiency of Large Language Models (LLMs) in coding tasks is often a reflection of their extensive pre-training corpora, which typically collapses when confronted with previo…
Bootstrapping Code Translation with Weighted Multilanguage Exploration
Yuhan Wu, Huan Zhang, Wei Cheng +3
Code translation across multiple programming languages is essential yet challenging due to two vital obstacles: scarcity of parallel data paired with executable test oracles, and o…
A Pair Programming Framework for Code Generation via Multi-Plan Exploration and Feedback-Driven Refinement
Huan Zhang, Wei Cheng, Yuhan Wu +1
Large language models (LLMs) have achieved impressive performance on code generation. Although prior studies enhanced LLMs with prompting techniques and code refinement, they still…