Astra reads and edits other models' weights without full retraining
Researchers say changing 0.016 percent of Qwen3.5-0.8B weights raised MATH-500 accuracy from 7.9 to 48.1 percent.
48
AI model
Gemma 3 270M
- Mentions · 7d1
- Mentions · 30d2
- All mentions2
- First seenOct 14, 2026
- Last mentioned5d ago
Timeline
2 articles · newest firstFrozen 16 MB codes replace learned embeddings in small models
A Gemma-shaped ternary model fits in a 46 MB file and runs 615–740 tokens per second on a laptop CPU.
52
Appears with
- Qwen3.5-0.8Bai model · 1
- llama.cppproduct · 1
- Astraai model · 1
- BitNet b1.58ai model · 1
Entities are extracted by the model from each article. First seen , last mentioned . Watching keeps it in this browser only (no account); the watchlist page and dashboard alerts use it.