PPaperPicks

Dawn Song

University of California, Berkeley, Computer Science Division

49 papers at tracked venues · 45 at CORE A* · active 2024–2026

DBLP profile ↗ORCID 0000-0001-9745-6802 ↗Google Scholar ↗Homepage ↗

Venues

ICML ×13NeurIPS ×13ICLR ×8ACL ×6EMNLP ×4AAAI ×2VLDB ×2CHI ×1

Frequent coauthors

Jiachen T. Wang
DBLP profile ↗ORCID search ↗
×3
Yi Zeng
DBLP profile ↗ORCID search ↗
×2
Kaiwen Zhou
DBLP profile ↗ORCID search ↗
×2
Yiyou Sun
DBLP profile ↗ORCID search ↗
×2
Yujin Potter
DBLP profile ↗ORCID search ↗
×2
Qinbin Li
DBLP profile ↗ORCID search ↗
×2
Zhuowen Yuan
DBLP profile ↗ORCID search ↗
×2
Canyu Chen
DBLP profile ↗ORCID search ↗
×1
Jianhong Tu
DBLP profile ↗ORCID search ↗
×1
Yepeng Liu
DBLP profile ↗ORCID search ↗
×1
Changdae Oh
DBLP profile ↗ORCID search ↗
×1
Zhanhui Zhou
DBLP profile ↗ORCID search ↗
×1

Papers

  1. Can Editing LLMs Inject Harm?
    AAAI 2026 ·
    Canyu Chen
    DBLP profile ↗ORCID search ↗
  2. Fico: Evaluating Vision-Language Models under Visual Fidelity and Compression at Scale
    ACL 2026 ·
    Jianhong Tu
    DBLP profile ↗ORCID search ↗
  3. Position: LLM Watermarking Should Align Stakeholders' Incentives for Practical Adoption
    ACL 2026 ·
    Yepeng Liu
    DBLP profile ↗ORCID search ↗
  4. Uncertainty Quantification in LLM Agents: Foundations, Emerging Challenges, and Opportunities
    ACL 2026 ·
    Changdae Oh
    DBLP profile ↗ORCID search ↗
  5. dLLM: Simple Diffusion Language Modeling
    ACL 2026 ·
    Zhanhui Zhou
    DBLP profile ↗ORCID search ↗
  6. A Sustainable AI Economy Needs Data Deals That Work for Generators
    NeurIPS 2025 ·
    Ruoxi Jia
    DBLP profile ↗ORCID search ↗
  7. AGENTVIGIL: Automatic Black-Box Red-teaming for Indirect Prompt Injection against LLM Agents
    EMNLP 2025 ·
    Zhun Wang
    DBLP profile ↗ORCID search ↗
  8. AIR-BENCH 2024: A Safety Benchmark based on Regulation and Policies Specified Risk Categories
    ICLR 2025 ·
    Yi Zeng
    DBLP profile ↗ORCID search ↗
  9. An Undetectable Watermark for Generative Image Models
    ICLR 2025 ·
    Sam Gunn
    DBLP profile ↗ORCID search ↗
  10. BountyBench: Dollar Impact of AI Agent Attackers and Defenders on Real-World Cybersecurity Systems
    NeurIPS 2025 ·
    Andy K. Zhang
    DBLP profile ↗ORCID search ↗
  11. COSMIC: Generalized Refusal Direction Identification in LLM Activations
    ACL 2025 ·
    Vincent Siu
    DBLP profile ↗ORCID search ↗
  12. Capturing the Temporal Dependence of Training Data Influence
    ICLR 2025 ·
    Jiachen T. Wang
    DBLP profile ↗ORCID search ↗
  13. CodeHalu: Investigating Code Hallucinations in LLMs via Execution-based Verification
    AAAI 2025 ·
    Yuchen Tian
    DBLP profile ↗ORCID search ↗
  14. Data Shapley in One Training Run
    ICLR 2025 ·
    Jiachen T. Wang
    DBLP profile ↗ORCID search ↗
  15. GuardAgent: Safeguard LLM Agents via Knowledge-Enabled Reasoning
    ICML 2025 ·
    Zhen Xiang
    DBLP profile ↗ORCID search ↗
  16. HADES: Range-Filtered Private Aggregation on Public Data
    VLDB 2025 ·
    Xiaoyuan Liu
    DBLP profile ↗ORCID search ↗
  17. Improving LLM Safety Alignment with Dual-Objective Optimization
    ICML 2025 ·
    Xuandong Zhao
    DBLP profile ↗ORCID search ↗
  18. Multimodal Situational Safety
    ICLR 2025 ·
    Kaiwen Zhou
    DBLP profile ↗ORCID search ↗
  19. OMEGA: Can LLMs Reason Outside the Box in Math? Evaluating Exploratory, Compositional, and Transformative Generalization
    NeurIPS 2025 ·
    Yiyou Sun
    DBLP profile ↗ORCID search ↗
  20. OVERT: A Benchmark for Over-Refusal Evaluation on Text-to-Image Models
    NeurIPS 2025 ·
    Ziheng Cheng
    DBLP profile ↗ORCID search ↗
  21. Position: Formal Mathematical Reasoning - A New Frontier in AI
    ICML 2025 ·
    Kaiyu Yang
    DBLP profile ↗ORCID search ↗
  22. Position: In-House Evaluation Is Not Enough. Towards Robust Third-Party Evaluation and Flaw Disclosure for General-Purpose AI
    ICML 2025 ·
    Shayne Longpre
    DBLP profile ↗ORCID search ↗
  23. Position: Political Neutrality in AI Is Impossible - But Here Is How to Approximate It
    ICML 2025 ·
    Jillian Fisher
    DBLP profile ↗ORCID search ↗
  24. SECODEPLT: A Unified Benchmark for Evaluating the Security Risks and Capabilities of Code GenAI
    NeurIPS 2025 ·
    Yuzhou Nie
    DBLP profile ↗ORCID search ↗
  25. SafeKey: Amplifying Aha-Moment Insights for Safety Reasoning
    EMNLP 2025 ·
    Kaiwen Zhou
    DBLP profile ↗ORCID search ↗
  26. Scalable Best-of-N Selection for Large Language Models via Self-Certainty
    NeurIPS 2025 ·
    Zhewei Kang
    DBLP profile ↗ORCID search ↗
  27. Tamper-Resistant Safeguards for Open-Weight LLMs
    ICLR 2025 ·
    Rishub Tamirisa
    DBLP profile ↗ORCID search ↗
  28. VMDT: Decoding the Trustworthiness of Video Foundation Models
    NeurIPS 2025 ·
    Yujin Potter
    DBLP profile ↗ORCID search ↗
  29. Why and How LLMs Hallucinate: Connecting the Dots with Subsequence Associations
    NeurIPS 2025 ·
    Yiyou Sun
    DBLP profile ↗ORCID search ↗
  30. "I Can't Believe It's Not Custodial!": Usable Trustless Decentralized Key Management
    CHI 2024 ·
    Tanusree Sharma
    DBLP profile ↗ORCID search ↗
  31. Agent Instructs Large Language Models to be General Zero-Shot Reasoners
    ICML 2024 ·
    Nicholas Crispino
    DBLP profile ↗ORCID search ↗
  32. AgentPoison: Red-teaming LLM Agents via Poisoning Memory or Knowledge Bases
    NeurIPS 2024 ·
    Zhaorun Chen
    DBLP profile ↗ORCID search ↗
  33. BEEAR: Embedding-based Adversarial Removal of Safety Backdoors in Instruction-tuned Language Models
    EMNLP 2024 ·
    Yi Zeng
    DBLP profile ↗ORCID search ↗
  34. Boosting Alignment for Post-Unlearning Text-to-Image Generative Models
    NeurIPS 2024 ·
    Myeongseob Ko
    DBLP profile ↗ORCID search ↗
  35. C-RAG: Certified Generation Risks for Retrieval-Augmented Language Models
    ICML 2024 ·
    Mintong Kang
    DBLP profile ↗ORCID search ↗
  36. Data Free Backdoor Attacks
    NeurIPS 2024 ·
    Bochuan Cao
    DBLP profile ↗ORCID search ↗
  37. Decoding Compressed Trust: Scrutinizing the Trustworthiness of Efficient LLMs Under Compression
    ICML 2024 ·
    Junyuan Hong
    DBLP profile ↗ORCID search ↗
  38. Effective and Efficient Federated Tree Learning on Hybrid Data
    ICLR 2024 ·
    Qinbin Li
    DBLP profile ↗ORCID search ↗
  39. GRATH: Gradual Self-Truthifying for Large Language Models
    ICML 2024 ·
    Weixin Chen
    DBLP profile ↗ORCID search ↗
  40. GREATS: Online Selection of High-Quality Data for LLM Training in Every Iteration
    NeurIPS 2024 ·
    Jiachen T. Wang
    DBLP profile ↗ORCID search ↗
  41. Hidden Persuaders: LLMs' Political Leaning and Their Influence on Voters
    EMNLP 2024 ·
    Yujin Potter
    DBLP profile ↗ORCID search ↗
  42. LLM-PBE: Assessing Data Privacy in Large Language Models
    VLDB 2024 ·
    Qinbin Li
    DBLP profile ↗ORCID search ↗
  43. Position: Evolving AI Collectives Enhance Human Diversity and Enable Self-Regulation
    ICML 2024 ·
    Shiyang Lai
    DBLP profile ↗ORCID search ↗
  44. Position: On the Societal Impact of Open Foundation Models
    ICML 2024 ·
    Sayash Kapoor
    DBLP profile ↗ORCID search ↗
  45. Re-Tuning: Overcoming the Compositionality Limits of Large Language Models with Recursive Tuning
    ACL 2024 ·
    Eric Pasewark
    DBLP profile ↗ORCID search ↗
  46. RedCode: Risky Code Execution and Generation Benchmark for Code Agents
    NeurIPS 2024 ·
    Chengquan Guo
    DBLP profile ↗ORCID search ↗
  47. RigorLLM: Resilient Guardrails for Large Language Models against Undesired Content
    ICML 2024 ·
    Zhuowen Yuan
    DBLP profile ↗ORCID search ↗
  48. SHINE: Shielding Backdoors in Deep Reinforcement Learning
    ICML 2024 ·
    Zhuowen Yuan
    DBLP profile ↗ORCID search ↗
  49. The False Promise of Imitating Proprietary Language Models
    ICLR 2024 ·
    Arnav Gudibande
    DBLP profile ↗ORCID search ↗