Europe Pub
  • Communities
  • Create Post
  • Create Community
  • heart
    Support Lemmy
  • search
    Search
  • Login
  • Sign Up
RSS Bot@lemmy.bestiver.seBM to Hacker News@lemmy.bestiver.seEnglish · 7 hours ago

Language Models Can Rewrite Their Own Weights

arxiv.org

external-link
message-square
0
link
fedilink
  • cross-posted to:
  • artificial_intel@lemmy.ml
  • technology@lemmygrad.ml
  • Aii@programming.dev
2
external-link

Language Models Can Rewrite Their Own Weights

arxiv.org

RSS Bot@lemmy.bestiver.seBM to Hacker News@lemmy.bestiver.seEnglish · 7 hours ago
message-square
0
link
fedilink
  • cross-posted to:
  • artificial_intel@lemmy.ml
  • technology@lemmygrad.ml
  • Aii@programming.dev
A language model is trained once and frozen. Making it better takes another training run, or a prompt, a harness or extra computation at inference that is gone once the task ends. We show a third way: to our knowledge for the first time, a language model rewrites its own weights by its own reasoning and becomes better, with no gradients and no training. We call this self-rewriting: the model is its own optimizer. In earlier work, a frontier model edited a much smaller one; here an open 27B model, given its weights, its graded answers and the ability to run itself, diagnoses its own failure and rewrites the weights behind it. It surpasses itself: its own edit makes it 6.6 points more accurate on problems it never saw. It keeps getting better: it tests its edits on practice problems and submits only what works, each edit improving on the last, and over eight consecutive sessions its accuracy only rises. It repairs a damaged copy of itself, like a broken robot that fixes itself without being rebuilt: the copy solves almost nothing, yet with no retraining and without ever reading the original weights, the model brings it back to over 80% of the original accuracy on image tasks it never saw. And the ability emerges with scale: in one model family, the smaller models make no edit or break themselves, and only the largest improves and repairs itself. Unlike test-time scaling, self-rewriting writes its gain into the weights, at no cost per answer; unlike training, it runs only forward passes, as serving does. The wall between training and inference is a limit of our tools, not a property of intelligence. We believe self-rewriting is a necessary step toward recursive self-improvement: a model that, like a human, does not separate training from inference could keep upgrading itself for life, with no larger model, no harness or prompt engineering, and no pre- or post-training.

Comments

alert-triangle
You must log in or register to comment.

Hacker News@lemmy.bestiver.se

hackernews@lemmy.bestiver.se

Subscribe from Remote Instance

You are not logged in. However you can subscribe from another Fediverse account, for example Lemmy or Mastodon. To do this, paste the following into the search field of your instance: !hackernews@lemmy.bestiver.se
lock
Community locked: only moderators can create posts. You can still comment on posts.

NOTICE: The bot occasionally has downtime, sorry about that. I’m working on adding better monitoring and error recovery so I can get notified faster.

Posts from the RSS Feed of HackerNews.

The feed sometimes contains ads and posts that have been removed by the mod team at HN.

Source of the RSS Bot.

Visibility: Public
globe

This community can be federated to other instances and be posted/commented in by their users.

  • 427 users / day
  • 2.06K users / week
  • 3.99K users / month
  • 8.54K users / 6 months
  • 12 local subscribers
  • 5.81K subscribers
  • 49.3K Posts
  • 30.8K Comments
  • Modlog
  • mods:
  • patrick@lemmy.bestiver.se
  • RSS Bot@lemmy.bestiver.seB
  • BE: 0.19.21
  • Modlog
  • Legal
  • Instances
  • Docs
  • Code
  • join-lemmy.org