jushengzhang/robot-weight-debugging
Resource
Can I trust it?
Worth a closer look before you use it.
Warnings Check
Very new: it was created 8 days ago. Wait a little before installing it.
Stars from real accounts No data
We haven't checked its stars for fakes yet.
Active OK
Last change 8 days ago.
Licence OK
Apache-2.0: you can use it for free, also in a commercial product. Keep the licence notice.
Quick checks from public data. They help, but they are no guarantee.
What it is
Research code and experiment logs for an approach where an LLM agent inspects a pretrained robot policy's weights and edits 65 token-embedding rows to repair instruction failures, using simulator feedback without gradient updates. Includes 16 editing-session transcripts and…
Who it's for
Robotics and machine learning researchers studying vision-language-action policies and model editing
How to use it
Ask your AI to set it up
Paste this into Claude, ChatGPT or your coding assistant.
Help me install and set up jushengzhang/robot-weight-debugging in my project. Follow its README: https://github.com/jushengzhang/robot-weight-debugging. Explain each step and check it is safe before running anything.
Star curve
Get alerts for repos like this
We'll email you when something in Models & inference starts rising. Free, once a week, only what matches.
What arrives: one email on Mondays with the repos that match, never more. See last week's. Privacy
For maintainers
Add this badge to your README. It shows the repo's place on the radar, updates every hour and links back to this page.
[](https://repoalerts.com/repo/jushengzhang/robot-weight-debugging?ref=badge)Similar repos
- NVlabs/ASPIRE
A research framework for robotics in which an LLM-driven coding agent refines sensorimotor skills through iterative debugging and evolutionary search instead of gradient-based training. Evaluated on L
- agentic-learning-ai-lab/deleting-weights
Research code for Deletion Tuning, a post-training method that learns a binary mask over the gate projection of an LLM's transformer blocks, zeroing a small fraction of pretrained weights. It covers t
- jushengzhang/CE-Graph
Reference implementation of a failure-driven loop that refines LLM workflows: it collects failed execution traces, clusters recurring failure modes, proposes prompt or graph edits, and accepts edits o
- wanghao-SunSky/cs329z-chinese
A Chinese-language self-study course on engineering AI agents, adapted from Stanford CS329Z. It covers LLMs, RAG, tool calling, memory, multi-agent collaboration, evaluation and safety across 19 lesso
- RikiyaT/lm_understand-logs
Complete logs of the experiments in a blog post on whether LLMs can understand and write neural network weights. Includes prompts, model replies, sandbox code and outputs, evaluations, and the network
- ZhuoyangLiu2005/agentic-wam
A research codebase for AW-0, a world action model for robot manipulation that jointly denoises a code plan, future video and robot actions in one mixture-of-transformers. It includes pretraining, pos