Log inSign up
Cunxiao Du
66 posts
@ducx_du

Cunxiao Du

@ducx_du
Research Scientist @ seed
Joined March 2024
192
Following
612
Followers
RepliesRepliesRepostsRepostsMediaMedia

Log in or sign up for X

See what’s happening and join the conversation

Continue with phone
or
Log in with username or email
Terms·Privacy·Cookies·Accessibility·Ads Info·© 2026 X Corp.
  • @ducx_du
    Cunxiao Du
    @ducx_du
    Jul 31
    Introducing AutoCompact: Learning When to Compact Context in Long-Horizon Coding Agents. Long-horizon coding trajectories often accumulate failed attempts, stale hypotheses, and noisy tool outputs, making it harder for the agent to focus on what matters. 👉AutoCompact teaches
    Image
    13
  • @ducx_du
    Cunxiao Du
    @ducx_du
    Jul 7
    Glad to see the ICML best paper: The Flexibility Trap: Rethinking the Value of Arbitrary Order in Diffusion Language Models also foucs on this question, and discuss our blog. I hope our community can study it further!
    Image
    @ducx_du
    Cunxiao Du
    @ducx_du
    Nov 25, 2025
    Diffusion LLMs (DLLM) can do “any-order” generation, in principle, more flexible than left-to-right (L2R) LLM. Our main finding is uncomfortable: ➡️ In real language, this flexibility backfires: DLLMs become worse probabilistic models than the L2R / R2L AR LMs. This
  • @ducx_du
    Cunxiao Du
    @ducx_du
    Dec 23, 2025
    robo phd @YouJiacheng is so professional on LLM🤣, I have never checked the annotation guidelines of OpenAI PRM800k, although I have read this paper so many times. But the idea did come from when @mavenlin and I were using codex, we just felt codex is so human-like. So we also
    @YouJiacheng
    You Jiacheng
    @YouJiacheng
    Dec 23, 2025
    This is similar to what OpenAI did in PRM800k (2023). I think this is the right way to collect data. github.com/openai/prm800k…
    Image
    1
  • @ducx_du
    Cunxiao Du
    @ducx_du
    Dec 18, 2025
    A simple and fast method for high-quality data annotation: On-Policy Annotation. Humans lightly edit LLM outputs, then let the LLM continue from the edited prefix—rather than labeling from scratch. Most tokens remain LLM-generated, boosting annotation efficiency and
    Image
    00:00
    1
  • @ducx_du
    Cunxiao Du
    @ducx_du
    Dec 17, 2025
    Can I say this is currently the most user-friendly agent framework I’ve tried? 🚀 You can use it directly in the terminal without Docker. At every step, you can fully see the model’s output and the commands to be executed and also freely edit them. All of this with minimal
    @LongxuDou
    Longxu Dou
    @LongxuDou
    Dec 17, 2025
    🚀We propose Reptile, a Terminal Agent🤖️that enables interaction with an LLM agent directly in your terminal. The agent can execute any command or custom CLI tool to accomplish tasks, and users can define their own tools and commands for the agent to utilize. ✨What Makes
    Image
Advertisement
Advertisement