Matryoshka NLA (Qwen2.5-7B L20) An NLA trained with random-length truncation of the verbalizer's output, so the most important information comes first. Checkpoints, data, baseline. syvb/nla-qwen2.5-7b-L20-v3-rl Updated 20 days ago syvb/nla-qwen2.5-7b-L20-av-matryoshka-sonnet46-v3 8B • Updated 20 days ago • 22 syvb/nla-qwen2.5-7b-L20-ar-matryoshka-sonnet46-v3 5B • Updated 20 days ago • 29 syvb/nla-qwen2.5-7b-L20-v3-rl-checkpoints Updated 20 days ago
NLA length penalty syvb/nanonla-qwen3-8b-L24-rl-lora Updated Jun 19 • 1 syvb/nanonla-qwen3-8b-L24-ar 5B • Updated Jun 19 • 17 • 1 syvb/nanonla-qwen3-8b-L24-av 8B • Updated Jun 19 • 47 • 1 syvb/nanonla-qwen3-8b-L24-results Viewer • Updated Jun 19 • 7k • 84 • 1
Matryoshka NLA (Qwen2.5-7B L20) An NLA trained with random-length truncation of the verbalizer's output, so the most important information comes first. Checkpoints, data, baseline. syvb/nla-qwen2.5-7b-L20-v3-rl Updated 20 days ago syvb/nla-qwen2.5-7b-L20-av-matryoshka-sonnet46-v3 8B • Updated 20 days ago • 22 syvb/nla-qwen2.5-7b-L20-ar-matryoshka-sonnet46-v3 5B • Updated 20 days ago • 29 syvb/nla-qwen2.5-7b-L20-v3-rl-checkpoints Updated 20 days ago
NLA length penalty syvb/nanonla-qwen3-8b-L24-rl-lora Updated Jun 19 • 1 syvb/nanonla-qwen3-8b-L24-ar 5B • Updated Jun 19 • 17 • 1 syvb/nanonla-qwen3-8b-L24-av 8B • Updated Jun 19 • 47 • 1 syvb/nanonla-qwen3-8b-L24-results Viewer • Updated Jun 19 • 7k • 84 • 1
Sleeping Agents NLA Qwen3.6-27B — Activation Explorer 🔬 Read Qwen3.6-27B activations in English (standard NLA)
Sleeping Agents NLA Qwen3.6-27B — Activation Explorer 🔬 Read Qwen3.6-27B's activations in plain English
Running on Zero Agents NLA Image Soft-Token Verbalizer (Gemma-3) 🔍 Verbalize Gemma-3 image soft tokens with the NLA AV