Unverified50% confidenceFactExact time
As a control, for the larger Gemma 3 model, which rarely loops, Astra could not improve its math score.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
11/4/2026
First Seen
Valid until: 2/2/2027
Sources
AI Directly Rewrites Model Weights: MIT Study Reveals New Black-Box Breakthrough
youtubeMatthew Berman11/4/2026
Related Entities
Related Claims
UnverifiedBy editing an extremely small portion of the weights, Astra raised QUEN 3.5 0.8b's solve rate on math problems from roughly 1 correct out of every 13 problems to close to half.61% similarUnverifiedIn experiment 8, Astra fixed the "looping" problem of the Qwen 3.5B (800 million parameter) model by editing its weights, after which the Math 500 score rose from slightly below 8% to slightly above 48% and GSM8K rose from 13.8% to 67.1%.60% similarUnverifiedAntAngelMed's INT4 quantized version requires only 2 Ascend 910B cards for inference.58% similarUnverifiedGPT-6 Sol achieves a slight lead over the previous generation Terra under the Max tier of the AA comprehensive capability leaderboard, but still trails Astra by a noticeable margin.57% similarUnverifiedAll three models (Sol, Luna, and DeepSeek) failed to accurately reproduce multi-image layouts, which was a common shortcoming.57% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1136678API
curl https://kongchang.com/api/v1/knowledge/claims/1136678MCP
get_claim(id=1136678)