HOT RADAR
llmHEAT 10.7°

AI CURATED HOT STORY · 11/21/2026

Stanford-led Study Finds 27B Open Model Can Self-Improve by Editing Its Own Weights

1 reports archived1 independent sourcesupdated 11/21/2026, 01:17:18
Synthesis & Latest Updates
1 Sources Cross-Validated

A new paper from Stanford and other top labs shows a 27B open model can improve itself without training or gradients: Qwen3.5-27B was given access to its own weights, graded practice answers and a Python session with gradients blocked, then reasoned over hidden states for completed versus truncated answers and weakened the direction separating them. On 490 held-out problems fixed in advance, accuracy rose from 57.2% to 63.8%, while the 2B model made no edit and the 4B and 9B models degraded themselves.

LATEST/A new paper from Stanford and other top labs shows a 27B open model can improve itself without training or gradients: Qwen3.5-27B was given access to its own weights, graded practice answers and a Python session with gradients blocked, then reasoned over hidden states for completed versus truncated answers and weakened the direction separating them. On 490 held-out problems fixed in advance, accuracy rose from 57.2% to 63.8%, while the 2B model made no edit and the 4B and 9B models degraded themselves.

Hourly heat curve

16 fully observed hours
Now
10.7
Peak
15.202:00
24h change
-30%

The line compares only sources observed throughout; gap hours are interpolated to keep the trend continuous.

Coverage timeline

Total 1 reports · Latest first
  1. Rohan Paul (@rohanpaul_ai)T2·78 pts
    • The model edits its own weights via reasoning with gradients blocked, bypassing training clusters and extra inference compute
    • Accuracy rose from 57.2% to 63.8% on 490 pre-registered held-out problems
    • There is a clear scale threshold: 2B produced no edit, while 4B and 9B degraded themselves
    • Practical guidance is to keep only self-edits that pass held-out tests