PS
Peiyang Song
cs.AIcs.LGcs.CLcs.LOcs.CYstat.MLcs.PLcs.SE
On Valency
published · living versionsW_pf26rsgq·v1 · currentpublished
Epi: Towards a Persistent, Self-Improving Agent for Autonomous Research
with Yue Huang, Zhangchen Xu, Xiaonan Luo, Zheyuan Liu +14
1 version
Preprints & journals
18 papers in the corpus · 2023–2026Epi: Towards a Persistent, Self-Improving Agent for Autonomous Research2610.24944v1 · Yue Huang, Zhangchen Xu, Xiaonan Luo et al.2026 · 0 citationsarXivon Valency
Representational Control over Self-Report & Behavior Coherence in LLM Risk-Taking2610.04125v1 · Rafal Kocielnik, Peiyang Song, Pengrui Han et al.2026 · 0 citationsarXiv
Vero: Can AI Agents Build Formally Verified Software Repositories?2608.13522v1 · Zhe Ye, Hantao Lou, Yuechun Sun et al.2026 · 0 citationsarXiv
P: Joint Program-and-Proof Planning for Verified Code Generation2608.09277v1 · Zenan Li, Ziran Yang, Peiyang Song et al.2026 · 0 citationsarXiv
Rethinking Psychometric Evaluation of LLMs: When and Why Self-Reports Predict Behavior2606.12730v1 · Rafal Kocielnik, Pengrui Han, Peiyang Song et al.2026 · 0 citationsarXiv
Human agency in initial human-AI proof formalization workflows2606.04273v2 · Katherine M. Collins, Simon Frieder, Jonas Bayer et al.2026 · 0 citationsarXiv
Interactive Evaluation Requires a Design Science2605.17829v1 · Keyang Xuan, Peiyang Song, Pan Lu et al.2026 · 0 citationsarXiv
Steer2Adapt: Dynamically Composing Steering Vectors Elicits Efficient Adaptation of LLMs2602.07276v1 · Pengrui Han, Xueqiang Xu, Keyang Xuan et al.2026 · 0 citationsarXiv
Large Language Model Reasoning Failures2602.06176v1 · Peiyang Song, Pengrui Han, Noah Goodman2026 · 3 citationsarXiv
How and Why LLMs Generalize: A Fine-Grained Analysis of LLM Reasoning from Cognitive Behaviors to Low-Level Patterns2512.24063v1 · Haoyue Bai, Yiyou Sun, Wenjie Hu et al.2025 · 0 citationsarXiv
Adaptation of Agentic AI: A Survey of Post-Training, Memory, and Skills2512.16301v3 · Pengcheng Jiang, Jiacheng Lin, Zhiyi Shi et al.2025 · 1 citationarXiv
The Personality Illusion: Revealing Dissociation Between Self-Reports & Behavior in LLMs2509.03730v2 · Pengrui Han, Rafal Kocielnik, Peiyang Song et al.2025 · 0 citationsarXiv
LeanProgress: Guiding Search for Neural Theorem Proving via Proof Progress Prediction2502.17925v3 · Robert Joseph George, Suozhi Huang, Peiyang Song et al.2025 · 0 citationsarXiv
LeanAgent: Lifelong Learning for Formal Theorem Proving2410.06209v8 · Adarsh Kumarappan, Mo Tiwari, Peiyang Song et al.2024 · 2 citationsarXiv
Creative and Context-Aware Translation of East Asian Idioms with GPT-42410.00988v1 · Kenan Tang, Peiyang Song, Yao Qin et al.2024 · 1 citationarXiv
In-Context Learning May Not Elicit Trustworthy Reasoning: A-Not-B Errors in Pretrained Language Models2409.15454v1 · Pengrui Han, Peiyang Song, Haofei Yu et al.2024 · 3 citationsarXiv
Lean Copilot: Large Language Models as Copilots for Theorem Proving in Lean2404.12534v3 · Peiyang Song, Kaiyu Yang, Anima Anandkumar2024 · 5 citationsarXiv
LeanDojo: Theorem Proving with Retrieval-Augmented Language Models2306.15626v2 · Kaiyu Yang, Aidan M. Swope, Alex Gu et al.2023 · 75 citationsarXiv
Career total: 21 works. 18 are in this corpus.
Profile built from the corpus for this byline.
Author records are still filling in while the Hub is in alpha. If this is your page, you'll be able to claim it soon. Spot a mistake? Tell us.