10 papers
Let the Library Speak: Self-Advertised Method Selection for Formal Proving
Xiaopeng Yuan, Suijin Wang, Yanli Wang +5
LLM-based formal provers can retrieve relevant lemmas and prior proofs, but relevance alone does not say whether a mathematical method can be used on the current theorem. A method…
Large-scale Repository Engineering via Agent-Native Reusable Code Primitives
Haibo Jin, Peng Kuang, Xucheng Yu +3
Large language models equipped with development environments have moved code generation toward repository-scale construction, yet building complete repositories remains difficult b…
Harness Engineering for Software Engineering via Modular Executable Dev-Primitives
Haibo Jin, Xinjie Li, Peng Kuang +1
Large language models (LLMs) equipped with terminal access have demonstrated strong capabilities in automating software engineering tasks. However, existing agents remain brittle o…
Composing Task-specific Agent Harnesses at Test Time with Reusable Primitives
Peng Kuang, Haibo Jin, Dehao Wu +4
Agent harnesses govern how large language models (LLMs) gather context, invoke tools, verify results, preserve state, and terminate, largely affecting agent performance. However, t…
ANTMAN: Adaptive Need Tracking for Multi-Agent Navigation in Large Information Spaces
Jerry Wang, Haibo Jin, Xiaopeng Yuan +2
Information-seeking agents increasingly operate over information spaces that are too large to process exhaustively. Yet many multi-agent systems organize computation around static…
BabelArena: A Large-Scale Multilingual Benchmark for LLM Agents
Peng Kuang, Yuchun Fan, Jiangnan Li +7
Large language model (LLM) agents increasingly execute multi-step workflows through tool use and interaction with users and environments. However, current agent evaluations are lar…
KV-PRM: Efficient Process Reward Modeling via KV-Cache Transfer for Multi-Agent Test-Time Scaling
Peng Kuang, Haibo Jin, Xiaoyu Han +5
Process Reward Models (PRMs) have been proven to be highly effective in guiding test-time scaling (TTS) methods, which significantly boost the capabilities of LLM-based multi-agent…
Closing the Loop on Latent Reasoning via Test-Time Reconstruction
Xiaopeng Yuan, Haibo Jin, Ye Yu +4
Recent work moves intermediate reasoning from natural-language traces into latent or cache-level representations to reduce token overhead and avoid a discrete communication bottlen…
GazePrinter: Visualizing Expert Gaze to Guide Novices in a New Codebase
Peng Kuang, Emma Söderberg, April Yi Wang +1
Program comprehension is an essential activity in software engineering. Not only does it often challenge professionals, but it can also hinder novices from advancing their programm…
Towards Real-world Debiasing: Rethinking Evaluation, Challenge, and Solution
Peng Kuang, Zhibo Wang, Zhixuan Chu +2
Spurious correlations in training data significantly hinder the generalization capability of machine learning models when faced with distribution shifts, leading to the proposition…