Last Updated: 2025-05-25
Time is the enemy.
About Me
Hey there! I’m Haohang Li, a Ph.D. student at the Cognition Lab and the Explainable and Controllable Artificial Intelligence Lab at Stevens Institute of Technology. I’m lucky to be advised by Prof. Jordan W. Suchow and Prof. Zining Zhu.
My research explores how language models (LMs) work—whether by studying their behavior or digging into their inner workings—and how we can make them more controllable and safer to use. If you’re working on anything related or just want to chat about LMs, feel free to email me at research at haohang.io or reach out on other places. I’m always open to collaborations or good conversations!
Research Interests
My research focuses are in two-fold:
Interpretability: I’m interested in understanding the behavior of language models through the lens of cognitive science, as well as from a mechanistically interpretable perspective. Gaining insight into how these models work is essential for building trust and ensuring their safe and scalable deployment.
Language Model Agents: I believe language model agents represent the future of how we interact with general-purpose foundation models. In my previous work, I’ve explored designing language model agents with memory mechanisms inspired by human cognitive structures, as well as developing multi-agent systems. Broadly, I’m interested in the cognitive design of agents and how they can communicate and collaborate to solve decision-making tasks or operate in open-ended environments.
Updates
[07/2025] Our paper FinMem: A Performance-Enhanced LLM Trading Agent with Layered Memory and Character Design has been accepted by IEEE Transaction on Big Data! It has been a long journey…
[05/2025] I received the “top reviewer” recognition from KDD 2025 Industry Track! Looking forward to see more great works.
[05/2025] We find strong evidence that language models have “truth neurons”! Check out our paper on arXiv!
[05/2025] Our paper INVESTORBENCH: A Benchmark for Financial Decision-Making Tasks with LLM-based Agent has been accepted by ACL Main!
[04/2025] I will join the research team @ Kensho for the summer as a research scientist intern!
[01/2025] Passed Qualifying Exam! Take a look at my responses if you’re interested in the process.
[09/2024] Two papers accepted by NeurIPS 2024!
Selected Publications
Please see my Google Scholar for the full list of my publications.
(★: First author / Co-first author)
★ <Pre-print> Truth Neurons: Haohang Li, Yupeng Cao, Yangyang Yu, Jordan W. Suchow, Zining Zhu.
★ <ACL 2025 Main> INVESTORBENCH: A Benchmark for Financial Decision-Making Tasks with LLM-based Agent: Haohang Li (*), Yupeng Cao (*), Yangyang Yu (*), Shashidhar Reddy Javaji, et al.
<NeurIPS 2024> FinBen: A Holistic Financial Benchmark for Large Language Models: Qianqian Xie, …, Haohang Li, et al.
★ <NeurIPS 2024> Fincon: A synthesized LLM multi-agent system with conceptual verbal reinforcement for enhanced financial decision making: Yangyang Yu (*), Zhiyuan Yao (*), Haohang Li (*), Zhiyang Deng (*), et al.
★ <LLM Agent @ ICLR 2024> FinMem: A Performance-Enhanced LLM Trading Agent with Layered Memory and Character Design: Haohang Li (*), Yangyang Yu (*), Zhi Chen (*), Yuechen Jiang (*), et al.
★ <Pre-print> TradingGPT: Multi-agent system with layered memory and distinct characters for enhanced financial trading performance: Yang Li (*), Yangyang Yu (*), Haohang Li (*), Zhi Chen (*), Khaldoun Khashanah.
Honors and Awards
- [2022] Fellowships Funded by Jefferies Financial Group.
- [2022 - 2026] Provost Fellowship.
- [2021] Outstanding Academic Achievement Award.
- [2019] Graduate Admission Scholarship.
Professional Services
Conference Reviewer: NeurIPS(2024), ICLR (2025), ICML(2025), COLM (2025), ACL ARR (2025), KDD (2024, 2025), ECML (2024).
Workshop / Challenge Organizer:
- FinNLP-AgentScen @ IJCAI2024.
- Agent-based Single Cryptocurrency Trading @ COLING-2025.
PC Member:
- FinNLP-FNP-LLMFinLegal @ COLING-2025.
- Workshop on Data Science @ INFORMS-2024.