Noah A. Smith
Professor
Researches natural language processing and machine learning, including open language models.
Noah A. Smith's lab at UW works on large language models. 98 papers since 2023, 50 at top venues, mostly NeurIPS and ICLR. 12 PhD students. 78 alumni, most often to Google and Amazon. Applications close .
At a glance
- papers, 2023–2026
- 26/ year
- at T1 venues
- 51%
- PhD students
- 12
- MS students
- 0
- alumni listed
- 78
Papers by yearfrom the lab homepage, 2023–2026
- 18 papers2026 · T1 8 · collecting
- 23 papers2025 · T1 12
- 32 papers2024 · T1 18
- 25 papers2023 · T1 12
T1 venuesOtherVenues: NeurIPS, ICLR, ACL, EMNLP, ICML
Research focuspaper titles that use each topic's words
- Large language models59 papers
- language53 papers
- open7 papers
- researches natural0 papers
- paul0 papers
Paper list
From the lab homepage.
202618 papers
- Proceedings of the IEEE Visualization and Visual Analytics ConferenceBeyond One Output: Visualizing and Comparing Distributions of Language Model Generations
- —Catching Fire in the News: The Necessary Conditions for Media Storms
- Proceedings of the Conference on Language ModelsEncode Once, Decode Never: Reusing Audio LM Internals for Efficient Temporal Localization
- Proceedings of the Conference on Language Models (COLMIn Proceedings of the Conference on Language Models (COLM 2026), San Francisco, CA, October 2026
- ICLR (T1)LEGATO: Large-scale End-to-end Generalizable Approach to Typeset OMR
- Proceedings of the Conference on Language ModelsMeta-Reinforcement Learning with Self-Reflection for Agentic Search
- Proceedings of the Conference on Language ModelsMMMG: A Comprehensive and Reliable Benchmark for Multitask Multimodal Generation
- Proceedings of the Conference on Language ModelsOlmo Hybrid: From Theory to Practice and Back
- Findings of the Annual Meeting of the Association for Computational LinguisticsReading Between the Lines: The One-Sided Conversation Problem
- Nature (T1)Retrofitting Language Models to Operate over Bytes
- ICLR (T1)RewardEval: Advancing Reward Model Evaluation
- TACL (T1)Rewriting History: A Recipe for Interventional Analyses to Study Data Effects on Model Behavior
202523 papers
- NeurIPS (T1)Broken Tokens? Your Language Model can Secretly Handle Non-Canonical Tokenizations
- Proceedings of the Conference of the North American Chapter of the Association for Computational LinguisticsComPO: Community Preferences for Language Model Personalization
- ICML (T1)DataDecide: How to Predict Best Pretraining Data with Small Experiments
- Proceedings of the Conference of the North American Chapter of the Association for Computational LinguisticsDoes Liking Yellow Imply Driving a School Bus? Semantic Leakage in Language Models
- Proceedings of the Conference on Language ModelsEstablishing Task Scaling Laws via Compute-Efficient Model Ladders
- CVPR (T1)Eval3D: Interpretable and Fine-grained Evaluation for 3D Generation
- NeurIPS (T1)FlexOLMo: Open Language Models for Flexible Data Use
- Proceedings of the Conference on Language ModelsFluid Language Model Benchmarking
- ACL (T1)Hybrid Preferences: Learning to Route Instances for Human vs. AI Feedback
- EMNLP (T1)Infini-gram mini: Exact n-gram Search at the Internet Scale with FM-Index
- Findings of the Annual Meeting of the Association for Computational LinguisticsLlamaPIE: Proactive In-Ear Conversation Assistants
- CVPR (T1)Molmo and PixMo: Open Weights and Open Data for State-of-the-Art Vision-Language Models
202432 papers
- Proceedings of the Joint International Conference on Computational Linguistics, Language Resources and EvaluationA Call for Clarity in Beam Search: How It Works and When It Stops
- ECCV (T1)BLINK: Multimodal Large Language Models Can See but Not Perceive
- EMNLP (T1)Breaking the Curse of Multilinguality with Cross-lingual Expert Language Models
- EMNLP WorkshopCPS-TaskForge: Generating Collaborative Problem Solving Environments for Diverse Communication Tasks
- NeurIPS (T1)Data Mixture Inference Attack: BPE Tokenizers Reveal Training Data Compositions
- NeurIPS (T1)Decoding-Time Language Model Alignment with Multiple Objectives
- Proceedings of the Conference on Language ModelsDoes Collaborative Human–LM Dialogue Generation Help Information Extraction from Human–Human Dialogues?
- ACL (T1)Dolma: an Open Corpus of Three Trillion Tokens for Language Model Pretraining Research
- Proceedings of the Conference on Causal Learning and ReasoningEstimating the Causal Effect of Early ArXiving on Paper Acceptance
- EMNLP (T1)Evaluating $n$-Gram Novelty of Language Models Using Rusty-DAWG
- Proceedings of the Neural Information Processing Systems Datasets and Benchmarks TrackEvaluating Copyright Takedown Methods for Language Models
- ICML (T1)How Language Model Hallucinations Can Snowball
202325 papers
- ICLR (T1)Binding Language Models in Symbolic Languages
- Findings of the Annual Meeting of the Association for Computational LinguisticsData-Efficient Finetuning Using Cross-Task Nearest Neighbors
- Findings of the Conference on Empirical Methods in Natural Language ProcessingDemystifying Prompts in Language Models via Perplexity Estimation
- EMNLP (T1)Do All Languages Cost the Same? Tokenization in the Era of Commercial Language Models
- ACL (T1)Elaboration-Generating Commonsense Question Answering at Scale
- NeurIPS (T1)Fine-Grained Human Feedback Gives Better Rewards for Language Model Training
- Proceedings of the Neural Information Processing Systems Datasets and Benchmarks TrackHow Far Can Camels Go? Exploring the State of Instruction Tuning on Open Resources
- Proceedings of the Joint Conference on Lexical and Computational SemanticsLEXPLAIN: Improving Model Explanations via Lexicon Supervision
- Findings of the Conference on Empirical Methods in Natural Language ProcessingMeasuring and Narrowing the Compositionality Gap in Language Models
- Journal of Natural Language EngineeringMorphosyntactic Probing of Multilingual BERT Models
- ACL (T1)NarrowBERT: Accelerating Masked Language Model Pretraining and Inference
- Findings of the Annual Meeting of the Association for Computational LinguisticsOne Embedder, Any Task: Instruction-Finetuned Text Embeddings
Where alumni went78 listed, 26% to industry
Industry 11Academia 31
Most often to
- Amazon
- Meta
Also to
- Emory University
- Hebrew University
- Hong Kong University
Apply to UWFall 2027
UW Allen School Computer Science PhD
- Round
- Single round
- Opens
- GRE
- not required
- English
- See program page
- Fee
- See program page
Not yet confirmed for this intake.
Similar labsnearest by research summary
- CAToronto
Zhijing Jin
- Large language models
- GBCambridge
Paula Buttery
- Large language models
- USUW
Hannaneh Hajishirzi
- Knowledge, QA & retrieval
- CNPeking
Zhi Jin
- Large language models
- CNPeking
Ge Li
- Large language models
- USUCSD
Taylor Berg-Kirkpatrick
- Large language models
Your notes
Save this lab to keep notes.
Data
Checked against the lab homepage.
- Collected
- Verified
- Papers
- The lab homepage
- Members
- Lab homepage
- Alumni
- Lab homepage
Tiers are venue classifications, not lab rankings.
UWcloses