Sean Welleck
Assistant Professor
Studies reasoning in language models, including mathematical reasoning and formal theorem proving.
Sean Welleck's lab at CMU works on reasoning & agents and large language models. 40 papers since 2023, 27 at top venues, mostly ICLR and NeurIPS. 7 PhD students and 5 master's students. Applications close .
At a glance
- papers, 2023–2026
- 10/ year
- at T1 venues
- 68%
- PhD students
- 7
- MS students
- 5
- alumni listed
- —
Papers by yearfrom the lab homepage, 2023–2026
- 16 papers2026 · T1 8 · collecting
- 11 papers2025 · T1 8
- 5 papers2024 · T1 4
- 8 papers2023 · T1 7
T1 venuesOtherVenues: ICLR, NeurIPS, EMNLP, ICML, ACL
Research focuspaper titles that use each topic's words
- Large language models13 papers
- Reasoning & agents7 papers
- proving3 papers
- theorem2 papers
- theorem proving1 paper
Paper list
From the lab homepage.
202616 papers
- —COLM 2026
- FMCADDSLean: A Framework for Type-Correct Interoperability Between Lean 4 and External DSLs (opens in a new tab)
- UIST (T1)Explorable Theorems: Making Written Theorems Explorable by Grounding Them in Formal Representations (opens in a new tab)
- COLMGradAlign: Gradient-Aligned Data Selection for LLM Reinforcement Learning (opens in a new tab)
- NeurIPS (T1)Gym-Anything: Turn any Software into an Agent Environment (opens in a new tab)
- —ICLR 2026
- ITPLeanArchitect: Automating Blueprint Generation for Humans and AI (opens in a new tab)
- COLMMind the Sim2Real Gap in User Simulation for Agentic Tasks (opens in a new tab)
- ICLR (Oral, (Top 1%))Premise Selection for a Lean Hammer (opens in a new tab)
- ICLR (T1)Programming with Pixels: Can Computer-Use Agents do Software Engineering? (opens in a new tab)
- ICML (T1)Propose, Solve, Verify: Self-Play Through Formal Verification (opens in a new tab)
- NeurIPS (T1)Reasoning with Latent Tokens in Diffusion Language Models (opens in a new tab)
202511 papers
- EMNLP (T1)Agentic-R1: Distilled Dual-Strategy Reasoning (opens in a new tab)
- ICML (T1)AlphaVerus: Bootstrapping Formally Verified Code Generation through Self-Improving Translation and Treefinement (opens in a new tab)
- ACL (T1)Evaluating Language Models as Synthetic Data Generators (opens in a new tab)
- ICLR (T1)ImProver: Agent-Based Automated Proof Optimization (opens in a new tab)
- ICLR (T1)Inference Scaling Laws: An Empirical Analysis of Compute-Optimal Inference for Problem-Solving with Language Models (opens in a new tab)
- COLML1: Controlling How Long A Reasoning Model Thinks With Reinforcement Learning (opens in a new tab)
- ICLR (Spotlight, (Top 5%))Lean-STaR: Learning to Interleave Thinking and Proving (opens in a new tab)
- ICLR (Oral, (Top 2%))miniCTX: Neural Theorem Proving with (Long-)Contexts (opens in a new tab)
- ICML (T1)Optimizing Temperature for Language Models with Multi-Sample Inference (opens in a new tab)
- EMNLP (T1)Rewarding the Unlikely: Lifting GRPO Beyond Distribution Sharpening (opens in a new tab)
- NAACL (T1)The BiGGen Bench: A Principled Benchmark for Fine-grained Evaluation of Language Models with Language Models (opens in a new tab)
20245 papers
- NeurIPS (T1)Easy-to-Hard Generalization: Scalable Alignment Beyond Human Supervision (opens in a new tab)
- TMLR (T1)From Decoding to Meta-Generation: Inference-time Algorithms for Large Language Models (opens in a new tab)
- ICLR (T1)Llemma: An Open Language Model for Mathematics (opens in a new tab)
- NeurIPS Safe Generative AIminiCodeProps: a Minimal Benchmark for Proving Code Properties (opens in a new tab)
- EMNLP (T1)Prometheus 2: An Open Source Language Model Specialized in Evaluating Other Language Models (opens in a new tab)
20238 papers
- ACL (T1)A Survey of Deep Learning for Mathematical Reasoning (opens in a new tab)
- ICLR (T1)Draft, Sketch, and Prove: Guiding Formal Theorem Provers with Informal Proofs (opens in a new tab)
- ICLR (T1)Generating Sequences by Learning to [Self-]Correct (opens in a new tab)
- EMNLP (T1)Inference-Time Policy Adaptors (opens in a new tab)
- NeurIPS (T1)Limits of Transformers on Compositionality (opens in a new tab)
- NeurIPS Math-AI WorkshopLLMstep: LLM proofstep suggestions in Lean (opens in a new tab)
- NeurIPS (T1)Self-refine: Iterative refinement with self-feedback (opens in a new tab)
- EMNLP (T1)STEER: Unified Style Transfer with Expert Reinforcement (opens in a new tab)
Where alumni went
Alumni are not listed on the lab homepage.
Apply to CMUFall 2027
CMU SCS Computer Science PhD
- Round
- Single round
- Opens
- GRE
- See program page
- English
- TOEFL / IELTS / Duolingo (F-1/J-1 non-native speakers)
- Fee
- USD 100 (USD 80 by early deadline)
Confirmed on the program page.
Similar labsnearest by research summary
- USUMD
Hal Daumé III
36 T1 - CNSJTU
Rui Wang
- Multilingual, translation & linguistics
- USCMU
William W. Cohen
- Large language models
- CNZhejiang
Xin Du
- Distributed, cloud & edge systems
- USCMU
Niloofar Mireshghallah
- Large language models
- USUCSD
Lianhui Qin
- Large language models
Your notes
Save this lab to keep notes.
Data
Some numbers could not be verified:
- The alumni list could not be found on the lab site.
- Collected
- Verified
- Papers
- The lab homepage
- Members
- Lab homepage
- Alumni
- Not collected
Tiers are venue classifications, not lab rankings.
CMUcloses