Swabha Swayamdipta
WiSE Gabilan Assistant Professor of Computer Science
Leads the DILL lab on datasets, interpretability and evaluation of language models.
Swabha Swayamdipta's lab at USC works on interpretability, evaluation & safety of LMs, large language models and trustworthy & safe ML. 35 papers since 2023, 16 at top venues, mostly ACL and Findings of EMNLP. 6 PhD students and 4 master's students. 13 alumni, most often to Google. Applications close .
At a glance
- papers, 2023–2026
- 9/ year
- at T1 venues
- 46%
- PhD students
- 6
- MS students
- 4
- alumni listed
- 13
Papers by yearfrom the lab homepage, 2023–2026
- 12 papers2026 · T1 7 · collecting
- 8 papers2025 · T1 1
- 10 papers2024 · T1 4
- 5 papers2023 · T1 4
T1 venuesOtherVenues: ACL, Findings of EMNLP, ICLR, EMNLP, Findings of ACL
Research focuspaper titles that use each topic's words
- Large language models20 papers
- Interpretability, evaluation & safety of LMs4 papers
- Trustworthy & safe ML0 papers
- datasets0 papers
- interpretability0 papers
Paper list
From the lab homepage.
202612 papers
- CHI (T1)Are We Automating the Joy Out of Work? Designing AI to Augment Work, Not Meaning (opens in a new tab)
- ACL (T1)Believing without Seeing: Quality Scores for Contextualizing Vision-Language Model Explanations (opens in a new tab)
- EMNLP (T1)BenchBrowser – Collecting Evidence for Evaluating Benchmark Validity (opens in a new tab)
- ICML (T1)Disentangling Geometry, Performance, and Training in Language Models (opens in a new tab)
- ICLR (T1)Every Language Model Has a Forgery-Resistant Signature (opens in a new tab)
- ICLR (T1)How Reliable is Language Model Micro-Benchmarking? (opens in a new tab)
- American Journal of Public HealthMortality Research With Artificial Intelligence Tools: Applying Language Models to the National Violent Death Reporting System Suicide Narratives, 2003–2023 (opens in a new tab)
- ICLR (T1)On the Trustworthiness of Generative Foundation Models: Guideline, Assessment, and Perspective (opens in a new tab)
- Proc. of COLMSample, Align, Synthesize: Graph-Based Response Synthesis with CONGRS (opens in a new tab)
- FAccTSide-by-side Comparison Amplifies Dialect Bias in Language Models (opens in a new tab)
- Proc. of ACL / NeurIPS Workshop on GenAI for Health / EAAMOUncovering Intervention Opportunities for Suicide Prevention with Language Model Assistants (opens in a new tab)
- Proc. of COLMWhy Fine-Tuning Encourages Hallucinations and How to Fix It (opens in a new tab)
20258 papers
- NeurIPS (T1)Better Language Model Inversion by Compactly Representing Next-Token Distributions (opens in a new tab)
- NeurIPS WorkshopChEmREF: Evaluating Language Model Readiness for Chemical Emergency Response Assistance (opens in a new tab)
- Findings of ACLELI-Why: Evaluating the Pedagogical Utility of LLM Explanations (opens in a new tab)
- Findings of EMNLPEvaluating Evaluation Metrics – The Mirage of Hallucination Detection (opens in a new tab)
- NeurIPS WorkshopEvaluation Under Imperfect Benchmarks and Ratings: A Case Study in Text Simplification (opens in a new tab)
- Findings of EMNLPImproving LLM Personas via Rationalization with Psychological Scaffolds (opens in a new tab)
- Findings of ACLRobust Data Watermarking in Language Models by Injecting Fictitious Knowledge (opens in a new tab)
- Conference on Language ModelingTeaching Models to Understand (but not Generate) High-risk Data (opens in a new tab)
202410 papers
- ACL (T1)Annotating FrameNet via Structure-Conditioned Language Generation (opens in a new tab)
- ICLR (T1)Closing the Curious Case of Neural Text Degeneration (opens in a new tab)
- Findings of EMNLPCompare without Despair: Reliable Preference Evaluation with Generation Separability (opens in a new tab)
- Proceedings of COLMCrowd-Calibrator: Can Annotator Disagreement Inform Calibration in Subjective Tasks? (opens in a new tab)
- ICASSP (T1)Does Video Summarization Require Videos? Quantifying the Effectiveness of Language in Video Summarization (opens in a new tab)
- Proceedings of VL/HCCGenerative Explanations for Program Synthesizers (opens in a new tab)
- Proceedings of COLMLogits of API-Protected LLMs Leak Proprietary Information (opens in a new tab)
- Findings of NAACLNeuroComparatives: Neuro-Symbolic Distillation of Comparative Knowledge (opens in a new tab)
- EMNLP (T1)OATH-Frames: Characterizing Online Attitudes Towards Homelessness via LLM Assistants (opens in a new tab)
- Findings of EMNLPOut-of-Distribution Detection through Soft Clustering with Non-Negative Kernel Regression (opens in a new tab)
20235 papers
- Findings of ACLCOBRA Frames: Contextual Reasoning about Effects and Harms of Offensive Statements (opens in a new tab)
- ACL (T1)I2D2: Inductive Knowledge Distillation with NeuroLogic and Self-Imitation (opens in a new tab)
- JMLR (T1)MAUVE Scores for Generative Models: Theory and Practice (opens in a new tab)
- ACL (T1)REV: Information-Theoretic Evaluation of Free-Text Rationales (opens in a new tab)
- EMNLP (T1)We’re Afraid Language Models Aren’t Modeling Ambiguity (opens in a new tab)
Where alumni went13 listed, 20% to industry
Industry 1Academia 4
Most often to
Also to
- Carnegie Mellon University
- UC Berkeley
Apply to USCFall 2027
Computer Science PhD (Viterbi)
- Round
- Single round (Dec 15)
- Opens
- —
- GRE
- See program page
- English
- See program page
- Fee
- See program page
Confirmed on the program page.
Similar labsnearest by research summary
- USNYU
He He
- Interpretability, evaluation & safety of LMs
- USUSC
Robin Jia
- Interpretability, evaluation & safety of LMs
- GBImperial
Marek Rei
- Language & society
- CNPeking
Huishuai Zhang
- Large language models
- CNZhejiang
Zhixuan Chu
- Large language models
- USUT Austin
Yizhong Wang
- 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.
USCcloses