- Query-Focused Retrieval Heads Improve Long-Context Reasoning and Re-ranking (Proceedings of EMNLP 2025)
- ChartMuseum: Testing Visual Reasoning Capabilities of Large Vision-Language Models (Proceedings of NeurIPS 2025 Datasets and Benchmarks Track)
- Linear Representation Transferability Hypothesis: Leveraging Small Models to Steer Large Models (Preprint)
- LongProc: Benchmarking Long-Context Language Models on Long Procedural Generation (Proceedings of COLM 2025)
- Understanding Synthetic Context Extension via Retrieval Heads (Proceedings of ICML 2025)
- To CoT or not to CoT? Chain-of-thought Helps Mainly on Math and Symbolic Reasoning (Proceedings of ICLR 2025)
- LoFiT: Localized Fine-tuning on LLM Representations (Proceedings of NeurIPS 2024)
- Linguistic Compression in Single-Sentence Human-Written Summaries (Findings of the Conference on Empirical Methods for Natural Language Processing (EMNLP), 2023)