{
  "$type": "site.standard.document",
  "bskyPostRef": {
    "cid": "bafyreifdmj3zrdrb7kkwblgpqqu2d7uszckt454wad35hohnzu3kqjzlpm",
    "uri": "at://did:plc:pgryn3ephfd2xgft23qokfzt/app.bsky.feed.post/3mmmjqke3m5c2"
  },
  "path": "/t/why-qwen-is-dynamically-stable-an-empirical-phase-map-of-10-llms/176177#post_3",
  "publishedAt": "2026-05-24T16:48:51.000Z",
  "site": "https://discuss.huggingface.co",
  "textContent": "That’s a fascinating observation, and it aligns perfectly with our geometric findings!\nIn our analysis, we found that Qwen models consistently occupy an ‘Adaptive’ dynamical regime (balanced flux and stability), whereas other models of similar size often drift into ‘Underactive’ (rigid) or ‘Chaotic’ (unstable) states.\nFrom a dynamics perspective, this ‘Adaptive’ state might be exactly what makes Qwen such a robust ‘student’ model:\nIt’s not too rigid: It has enough internal flexibility (flux) to adapt to new instructions during fine-tuning without breaking its core structure.\nIt’s not too chaotic: It maintains enough stability to retain its original capabilities and avoid catastrophic forgetting.\nEssentially, its hidden-state trajectory is ‘resilient’ rather than ‘fragile.’ It would be interesting to see if this dynamic signature persists in the newer Qwen-3B or 7B versions after heavy fine-tuning. Have you noticed any specific prompts where Qwen feels too stable or too flexible?\"",
  "title": "Why Qwen is Dynamically Stable: An Empirical Phase Map of 10 LLMs"
}