Abstract
Language models are trained once, frozen and deployed: improving one again takes another training run, and a prompt, a scaffold or extra computation at inference is gone once the task ends. A frontier model has been shown to edit the weights of a much smaller one. We ask whether a model can rewrite its own weights to make itself better, and find that it can, by its own reasoning, with no gradients and no training; we call this self-rewriting. Given its weights, graded answers and the ability to run itself, a 27B model diagnoses its failure and edits the weights behind it. It surpasses itself: its own edit makes it 6.6 points more accurate on 490 problems it never saw, in a test fixed in advance, and another session's edit also makes it significantly better. It keeps getting better: it submits an edit only when its own tests support it, each edit replacing the last, and over its first eight such sessions its accuracy only rises. Unlike test-time scaling, the gain persists in its weights and adds no cost per answer; unlike training, it needs no pre- or post-training infrastructure, only an open model and hours on one GPU. It repairs a damaged copy of itself: it locates the corrupted matrix and rewrites it with values of its own, never reading the undamaged weights; found on text math problems alone, the repair lifts the copy on six image benchmarks it never saw from 7% to 70% accuracy, 83% of the undamaged model's. And the ability emerges with scale: across four sizes of one family, the 2B model makes no edit and the 4B and 9B models break themselves; only the 27B model works on itself as a human would, comparing its hidden states on answers that finish and on answers that run out of tokens, weakening the direction that separates them. Like a human, a model need not separate training from inference; improving a language model may come to take a conversation rather than a training run.
Community
A language model rewrites its own weights by its own reasoning and gets better, with no gradients and no training. It diagnoses its own failures, keeps improving across sessions, and repairs a damaged copy of itself. Demos and videos: https://www.jiajunfan.com/projects/self-rewriting/
This is an automated message from the Librarian Bot. I found the following papers similar to this paper.
The following papers were recommended by the Semantic Scholar API
- Inference for Reasoning Models: Practical Lessons from Evaluation and Serving (2026)
- IWIBench: Evaluating Gradient-Free Model Driven Intra-weight Injection (2026)
- Weight Space Programming for Vision Language Model Repair (2026)
- Can Language Models Learn While They Think? Testing Gradient Adaptation Across Architectures (2026)
- Rewriting Scaling Laws for Synthetic Data with Compute and Data Constraints (2026)
- VERS: Steering Reasoning Models by Editing Their Activation Explanations (2026)
- There Is No Free Loop: When Does Looping a Trained Model Pay for Its Inference Compute? (2026)
Please give a thumbs up to this comment if you found it helpful!
If you want recommendations for any Paper on Hugging Face checkout this Space
You can directly ask Librarian Bot for paper recommendations by tagging it in a comment: @librarian-bot recommend
Get this paper in your agent:
hf papers read 2611.21885 Don't have the latest CLI?
curl -LsSf https://hf.co/cli/install.sh | bash Models citing this paper 0
No model linking this paper
Datasets citing this paper 0
No dataset linking this paper
Spaces citing this paper 0
No Space linking this paper
Collections including this paper 0
No Collection including this paper