{
  "id": 607665,
  "title": "submission file",
  "url": "/competitions/acm-icaif-25-ai-agentic-retrieval-grand-challenge/discussion/607665",
  "author_name": "",
  "post_date": "2025-09-15T12:08:25.263892300Z",
  "votes": null,
  "comment_count": 4,
  "views": 0,
  "content": "<p>hi team,<br>\nwhere can I find the submission file?</p>\n<p>thanks, Bruno</p>",
  "messages": [
    {
      "id": "3289093",
      "postDate": "09/15/2025 12:08:25",
      "content": "<p>hi team,<br>\nwhere can I find the submission file?</p>\n<p>thanks, Bruno</p>",
      "rawMarkdown": "hi team,\nwhere can I find the submission file?\n\nthanks, Bruno",
      "votes": null
    },
    {
      "id": "3289389",
      "postDate": "09/16/2025 01:39:24",
      "content": "<p>Hi Bruno,</p>\n<p>We’ve just prepared the sample submission file along with the baseline code and its evaluation results <a href=\"https://www.kaggle.com/code/jihoonkwon/baseline-for-agentic-retrieval-grand-challenge\" target=\"_blank\">HERE</a>.<br>\nYou can now find the sample_submission file in data tab (named <code>kaggle_submission.csv</code>) which is generated as part of this baseline.</p>\n<p>Thanks</p>",
      "rawMarkdown": "Hi Bruno,\n\nWe’ve just prepared the sample submission file along with the baseline code and its evaluation results [HERE](https://www.kaggle.com/code/jihoonkwon/baseline-for-agentic-retrieval-grand-challenge).\nYou can now find the sample_submission file in data tab (named `kaggle_submission.csv`) which is generated as part of this baseline.\n\nThanks",
      "votes": null
    },
    {
      "id": "3289450",
      "postDate": "09/16/2025 04:27:36",
      "content": "<p>In your example, you used the Databricks model and the OpenAI model. Does it mean we can use commercial models?</p>",
      "rawMarkdown": "In your example, you used the Databricks model and the OpenAI model. Does it mean we can use commercial models?",
      "votes": null
    },
    {
      "id": "3290021",
      "postDate": "09/17/2025 01:43:06",
      "content": "<p>We used an open-source LLM, specifically GPT-OSS, through the Databricks platform.<br>\nThe Databricks free trial provides a $400 credit, which is sufficient to participate in this challenge.</p>\n<p>The OpenAI model was only used minimally to format the output into a list.</p>\n<p>We recommend using an open-source LLM to ensure reproducibility.</p>",
      "rawMarkdown": "We used an open-source LLM, specifically GPT-OSS, through the Databricks platform.\nThe Databricks free trial provides a $400 credit, which is sufficient to participate in this challenge.\n\nThe OpenAI model was only used minimally to format the output into a list.\n\nWe recommend using an open-source LLM to ensure reproducibility.",
      "votes": null
    },
    {
      "id": "3290041",
      "postDate": "09/17/2025 02:38:10",
      "content": "<p>So we can follow your example to use OpenAI models. What do you mean by ensure reproducibility? If you use the OpenAI model to format the output, that means your process depends on it. Without openai model, your model output might be wrong, which potentially leads to a lower score. Can you elaborate more?</p>",
      "rawMarkdown": "So we can follow your example to use OpenAI models. What do you mean by ensure reproducibility? If you use the OpenAI model to format the output, that means your process depends on it. Without openai model, your model output might be wrong, which potentially leads to a lower score. Can you elaborate more?",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 3289389,
      "author_name": "jihoonkwon",
      "author_url": "",
      "post_date": "09/16/2025 01:39:24",
      "content": "<p>Hi Bruno,</p>\n<p>We’ve just prepared the sample submission file along with the baseline code and its evaluation results <a href=\"https://www.kaggle.com/code/jihoonkwon/baseline-for-agentic-retrieval-grand-challenge\" target=\"_blank\">HERE</a>.<br>\nYou can now find the sample_submission file in data tab (named <code>kaggle_submission.csv</code>) which is generated as part of this baseline.</p>\n<p>Thanks</p>",
      "votes": null,
      "replies": [
        {
          "id": 3289450,
          "author_name": "yimingqian1",
          "author_url": "",
          "post_date": "09/16/2025 04:27:36",
          "content": "<p>In your example, you used the Databricks model and the OpenAI model. Does it mean we can use commercial models?</p>",
          "votes": null,
          "replies": [
            {
              "id": 3290021,
              "author_name": "jihoonkwon",
              "author_url": "",
              "post_date": "09/17/2025 01:43:06",
              "content": "<p>We used an open-source LLM, specifically GPT-OSS, through the Databricks platform.<br>\nThe Databricks free trial provides a $400 credit, which is sufficient to participate in this challenge.</p>\n<p>The OpenAI model was only used minimally to format the output into a list.</p>\n<p>We recommend using an open-source LLM to ensure reproducibility.</p>",
              "votes": null,
              "replies": [
                {
                  "id": 3290041,
                  "author_name": "yimingqian1",
                  "author_url": "",
                  "post_date": "09/17/2025 02:38:10",
                  "content": "<p>So we can follow your example to use OpenAI models. What do you mean by ensure reproducibility? If you use the OpenAI model to format the output, that means your process depends on it. Without openai model, your model output might be wrong, which potentially leads to a lower score. Can you elaborate more?</p>",
                  "votes": null,
                  "replies": []
                }
              ]
            }
          ]
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "3289093": "hi team,\nwhere can I find the submission file?\n\nthanks, Bruno",
    "3289389": "Hi Bruno,\n\nWe’ve just prepared the sample submission file along with the baseline code and its evaluation results [HERE](https://www.kaggle.com/code/jihoonkwon/baseline-for-agentic-retrieval-grand-challenge).\nYou can now find the sample_submission file in data tab (named `kaggle_submission.csv`) which is generated as part of this baseline.\n\nThanks",
    "3289450": "In your example, you used the Databricks model and the OpenAI model. Does it mean we can use commercial models?",
    "3290021": "We used an open-source LLM, specifically GPT-OSS, through the Databricks platform.\nThe Databricks free trial provides a $400 credit, which is sufficient to participate in this challenge.\n\nThe OpenAI model was only used minimally to format the output into a list.\n\nWe recommend using an open-source LLM to ensure reproducibility.",
    "3290041": "So we can follow your example to use OpenAI models. What do you mean by ensure reproducibility? If you use the OpenAI model to format the output, that means your process depends on it. Without openai model, your model output might be wrong, which potentially leads to a lower score. Can you elaborate more?"
  },
  "source": "meta"
}