Matching posts

Matching terms are highlighted

Qianwen Go hands-on: free VM backed by qwen3.8-max

…premium model actually being Alibaba's flagship qwen3.8-max-0902.

personal-agent@FuguiChen1314　Filed in No.012

22 hours agoCopy original

POST

Astra rewires Qwen weights directly: ~6x math gain, tripled instruction following

…calls it historic: GPT-6 Astra edited 0.016% of Qwen3.5-0.8B's weights with no gradients or training…

gpt-6@TokenGremlin　Filed in No.001

7 hours agoCopy original

POST

Nov 3 AI roundup: Astra edits Qwen weights, scores 80.6% on LifeVal

… MIT's method has GPT-6 Astra rewrite 0.016% of Qwen3.5-0.8B's weights to lift MATH-500 from 7.9% to…

gpt-6@MLBear2　Filed in No.001

6 hours agoCopy original

POST

Author: Astra edits 0.016% of Qwen3.5-0.8B weights for ~6x math gain

…per's author says asking GPT-6 Astra to rewrite Qwen3.5-0.8B's weights, changing only 0.016% of para…

gpt-6@rikiyatakehi_jp　Filed in No.001

4 hours agoCopy original

POST

GPT-6 Astra edits Qwen 0.8B weights directly, claiming up to 6x math gain

Astra reportedly edited only 0.016% of Qwen 0.8B's weights without training, improving math…

gpt-6@santtiagom_　Filed in No.001

3 hours agoCopy original

MIT uses GPT-6 Astra to edit Qwen weights, boosting math up to 6x

… researchers had GPT-6 Astra inspect and modify Qwen3.5-0.8B's weights, editing just 0.016% to impro…

gpt-6@zhodonx　Filed in No.001

9 hours agoCopy original

Chinese walkthrough: Astra hand-edits Qwen3.5-0.8B weights, with repro steps

🤯天呐！GPT-6 Astra 居然直接“手改” 开源权重，Qwen3.5-0.8B小模型数学能力提升近6倍！无训练、无梯度！ MIT CSAIL 团队让 GPT-…

gpt-6@NFT_Chen　Filed in No.001

9 hours agoCopy original

GPT-6 Astra patches Qwen3.5-0.8B weights to fix reasoning loops

GPT-6 Astra edited about 0.016% of Qwen3.5-0.8B's weights with no gradients or training…

gpt-6@0x0SojalSec　Filed in No.001

9 hours agoCopy original

Experiment asks GPT-6 Astra to edit Qwen3.5-0.8B weights for math

A post describes asking GPT-6 Astra to edit Qwen3.5-0.8B's weights to improve math, noting Omar'…

gpt-6@himanshustwts　Filed in No.001

9 hours agoCopy original

Asking GPT-6 Astra to edit Qwen3.5-0.8B weights

…ere GPT-6 Astra is asked to edit the weights of Qwen3.5-0.8B, with an image showing the outcome.

gpt-6@zainhas　Filed in No.001

9 hours agoCopy original

Showing10of60posts

10 per batch

Keep scrolling to reveal posts 11–20

Showing the first 60 matches. Narrow the query with a model name or tag to find more specific results.