2601.21476v1.pdf

{
  "file/base": {
    "record": {
      "extension": ".pdf",
      "hash": "5269c2b45b7d934d7c94f899a76d1f950224a9a78ba6d4f4065526c0aa76ea87",
      "media_type": "application/pdf",
      "media_type_prefix": "application",
      "name": "2601.21476v1.pdf",
      "size": 1147501
    },
    "source": {
      "execution_id": "469c0938-7a70-4618-8662-4f2444a8cfbe",
      "id": "file/base",
      "type": "Model",
      "version": "1.0.0"
    }
  },
  "file/pdf": {
    "private": false,
    "record": {
      "keywords": [],
      "page_count": 13,
      "producer": "pikepdf 8.15.1",
      "title": "SOUP: Token-level Single-sample Mix-policy Reinforcement Learning for Large Language Models",
      "version": "1.7"
    },
    "source": {
      "execution_id": "ef5392d4-2892-4e27-9798-cb1c19ac9df2",
      "id": "dorsal/pdf",
      "type": "Model",
      "version": "1.1.0"
    }
  }
}