Scholar
Lisa Dunlap
Google Scholar ID: Vy16O5UAAAAJ
PhD Student, University of California Berkeley
Citations & Impact
Citations
1,310
H-index
13
i10-index
14
Publications
20
Co-authors
0
Publications
Resume (English only)
Academic Achievements
- Discovering Divergent Representations between Text-to-Image Models (ICCV 2025)
- Video Action Differencing (ICLR 2025)
- VisionArena: 230K Real World User-VLM Conversations with Preference Labels (CVPR 2025)
- VibeCheck: Discover and Quantify Qualitative Differences in Large Language Models (ICLR 2025)
- From Crowdsourced Data to High-Quality Benchmarks: Arena-Hard and BenchBuilder Pipeline (ICML 2025)
- Describing Differences in Image Sets with Natural Language (CVPR 2024 oral)
- See, Say, and Segment: Teaching LMMs to Overcome False Premises (CVPR 2024)
- Diversify Your Vision Datasets with Automatic Diffusion-Based Augmentation (NeurIPS 2023)
- Diffusion Hyperfeatures: Searching Through Time and Space for Semantic Correspondence (NeurIPS 2023)
Research Experience
- Core contributor to Chatbot Arena, a community-driven platform for evaluating these models in the wild. Collaborated with other researchers including Ion Stoica, Serena Yeung-Levy, and more.
Education
- PhD student at UC Berkeley; Advisors: Joey Gonzalez (Sky Computing Lab), Trevor Darrell, and Jacob Steinhardt (Berkeley Artificial Intelligence Research lab).
Background
- Research interests: how data shapes model behavior; Specialties: building systems for extracting insights from model predictions, analyzing large-scale text and image datasets, and improving data quality for ML pipelines.
Miscellany
- Personal note: Has Tourette Syndrome, so please excuse any snorting, squeaking, eye rolling, or other out-of-distribution behavior if you meet her IRL. Wrote an essay on the topic.
Co-authors