2601.06142v1.pdf

{
  "dorsal/arxiv": [
    {
      "date_modified": "2026-02-17T05:53:05.099000Z",
      "hash": "1aa4740d88f5a8c41fac401b4a57b79b303d1f5827f58ed2531069d20a5a7441",
      "id": "575afba9-0d14-4d12-90f5-bc82a006e115",
      "source": {
        "execution_id": "8ee77e1c-294d-48eb-98f6-cbfedcdd427f",
        "id": "arXiv Dataset",
        "type": "Model",
        "variant": "snapshot-2026-01-17",
        "version": "0.1.0"
      },
      "user_id": 1000002
    }
  ],
  "file/base": {
    "record": {
      "extension": ".pdf",
      "hash": "1aa4740d88f5a8c41fac401b4a57b79b303d1f5827f58ed2531069d20a5a7441",
      "media_type": "application/pdf",
      "media_type_prefix": "application",
      "name": "2601.06142v1.pdf",
      "size": 608090
    },
    "source": {
      "execution_id": "450b3d8f-f806-4ce4-a04c-646d146d0146",
      "id": "file/base",
      "type": "Model",
      "version": "1.0.0"
    }
  },
  "file/pdf": {
    "private": false,
    "record": {
      "keywords": [],
      "page_count": 9,
      "producer": "pikepdf 8.15.1",
      "title": "Is Sanskrit the most token-efficient language? A quantitative study using GPT, Gemini, and SentencePiece",
      "version": "1.7"
    },
    "source": {
      "execution_id": "aabf1aef-d38d-4651-8051-3da04ba33ae4",
      "id": "dorsal/pdf",
      "type": "Model",
      "version": "1.1.0"
    }
  }
}