Hi there!

I am Xiaotian Han. I am training Large Language Models (LLMs) on the Rufus team at Amazon.

google scholar / 𝕏

🔥 News

• 2025.09: Four papers accepted by NeurIPS2025! [Longer Context Deeper Thinking] [NeuralPLexer3] [Capsule Prompt Tuning] [Grammars of Formal Uncertainty]. See you in San Diego!

• 2025.08: Four papers accepted by EMNLP2025! [MEPT] [Truthful Representations Flip] [Pruning Weights but Not Truth] [Quantized but Deceptive?]

• 2025.07: One paper on Speculative Thinking is accepted by COLM2025!

• 2025.06: I developed a Typst-based homepage template and migrated my website to it! Check out github repo!

• 2025.05: New preprints on Speculative Thinking, SELF, and Longer Context, Deeper Thinking form my group are out!

• 2025.05: Two papers on 💯-LongBench and CausalRAG accepted by ACL2025 Findings!

• 2025.04: Invited to serve as an Area Chair for NeurIPS2025!

• 2025.02: New preprint on Thinking Preference Optimization is out!

• 2025.02: One paper on You Only Debias Once is accepted by CPAL 2025 (Proceedings Track) Oral!

• More…

  • 2024.12: Technical report of NeuralPLexer3 for biomolecular complex structure prediction is out!
  • 2024.11: Graph Convolution≈Mixup is selected as Oral Presentation at LoG2024 TMLR Track!
  • 2024.08: One paper on Understanding Graph Convolution accepted by TMLR!
  • 2024.08: Excited to join Case Western Reserve University as an assistant professor!
  • 2024.07: I am excited to start my MLsys internship at Iambic Therapeutics

  • 2024.06: LLM Maybe LongLM has been selected as Spotlight (3.5%) at ICML2024!

    • 2024.05: One paper is accepted by ACL2024!
    • 2024.04: LLM Maybe LongLM: Self-Extend LLM Context Window Without Tuning is accepted by ICML2024!
    • 2024.05: One paper is accepted by ACL2024!
    • 2024.04: LLM Maybe LongLM: Self-Extend LLM Context Window Without Tuning is accepted by ICML2024!
    • 2024.04: My LiteLLaMa has been downloaded over 230K times on HuggingFace!
    • 2024.04: Honored to receive the Jane Street Graduate Research Fellowship Award Honorable Mention.
    • 2024.03: Implemented Triton based flash self-extend. Please try FlashSelfExtend!
    • 2024.01: One paper on Fairness Benchmark accepted by ICLR2024!
    • 2024.01: Our Survey on LLMs accepted by TKDD!
    • 2024.01: New preprint LLM Maybe LongLM: Self-Extend LLM Context Window Without Tuning!
    • 2023.12: One paper accepted by AAAI2024-SRRAI.
    • 2023.09: One paper accepted by NeurIPS2023.
    • 2023.07: 🔥🔥 Thrilled to release my LiteLLaMa on HuggingFace, try it out!
    • 2023.07: Our paper LLM for Clinical Text Mining accepted by AMIA2023!
    • 2023.05: One paper accepted by TMLR, Retiring ∆DP!
    • 2023.05: Thrilled to start my internship at Amazon.
    • 2023.01: One paper accepted by ICLR2023, MLPInit.
    • 2022.09: Thrilled to start my internship at Meta, work with Qifan Wang.
    • 2022.07: Our Paper $𝒢$-Mixup is awarded an Outstanding Paper Award at ICML 2022!
    • 2022.05: Thrilled to start my internship at Snap Inc., work with Neil Shah.
    • 2022.05: One paper accepted by ICML2022 (Oral).
    • 2022.01: One paper accepted by ICLR2022.
    • 2022.01: One paper accepted by TheWebConf2022.
    • 2020.05: One paper accepted by RecSys2020.