episode_id large_stringlengths 32 32 | gen large_stringclasses 1
value | position int64 0 0 | ts large_stringlengths 32 32 | dataset_id large_stringlengths 12 73 | mode large_stringclasses 1
value | backend large_stringclasses 1
value | model large_stringclasses 1
value | revision large_stringclasses 1
value | temperature float64 0 0 | tldr large_stringlengths 0 422 | terminal large_stringclasses 3
values | trajectory large_stringlengths 748 15.8k | card large_stringclasses 18
values | reasonings large_stringlengths 2 12.6k | observations large_stringlengths 2 12.9k | screens large_stringlengths 160 2.28k | n_turns int64 1 7 | coverage large_stringclasses 1
value | rating large_stringclasses 3
values | chosen_gen large_stringclasses 1
value | rated_ts large_stringclasses 15
values | verdict large_stringclasses 2
values |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
31acd22e6f6743a78aaf3255d2b2a40b | a | 0 | 2026-08-06T06:58:46.532025+00:00 | CyanMonkey/Dataset-2 | single | endpoint | LiquidAI/LFM2.5-2.6B | main | 0 | This dataset is a mixed pretraining corpus containing 500,000 rows designed for language model training, providing diverse text data across multiple domains. | submit | [{"role": "user", "content": "Write ONE sentence (max 50 words, English) saying what this dataset is and what it is\nfor: domain, data source/provenance, method, languages, distinctive capability when stated.\nDo not restate license, popularity, or usage instructions. Do not invent anything \u2014 every claim\nmust com... | # Dataset
Mixed pretraining corpus - 500,000 rows, 1 shards. | ["The user wants me to write a single sentence (max 50 words) describing what this dataset is, including its domain, data source/provenance, method, languages, and distinctive capabilities. I need to use the tools to get information about this dataset before providing the TLDR.\n\nLet me start by calling the splits fun... | [{"obs_index": 0, "tool": "splits", "op": "", "content": "{\"splits\": [{\"dataset\": \"CyanMonkey/Dataset-2\", \"config\": \"default\", \"split\": \"train\"}], \"pending\": [], \"failed\": []}", "coverage": ""}, {"obs_index": 1, "tool": "statistics", "op": "", "content": "{\"num_examples\": 500000, \"statistics\": [{\... | {"verdict": "pass", "refusal": false, "format": {"pass": true, "words": 22, "one_sentence": true}, "meta_language": {"pass": true, "matches": []}, "numbers": [{"text": "500,000", "normalized": "500000", "badge": "ok", "sources": [{"obs_index": 1, "tool": "statistics", "op": null, "partial": false}, {"obs_index": null, ... | 4 | pass | ||||
e47223c3e66443bc806b119c3c208139 | a | 0 | 2026-08-07T15:46:40.303953+00:00 | laion/dev_set_v2_a1_magicoder_20260805_125433 | single | endpoint | LiquidAI/LFM2.5-2.6B | main | 0 | The dataset `laion/dev_set_v2_a1_magicoder_20260805_125433` is a Laion-derived code training set containing 3,293 examples with agent-labeled conversations and timestamped entries, primarily focused on magic programming tasks using Python-like language constructs. | prose_fallback | [{"role": "user", "content": "Write ONE sentence (max 50 words, English) saying what this dataset is and what it is\nfor: domain, data source/provenance, method, languages, distinctive capability when stated.\nDo not restate license, popularity, or usage instructions. Do not invent anything \u2014 every claim\nmust com... | ["The user wants me to write a single sentence (max 50 words) describing what this dataset is - its domain, data source/provenance, method, languages, and distinctive capabilities. I need to use the available tools to gather information about the dataset before making my statement.\n\nLet me start by checking the split... | [{"obs_index": 0, "tool": "splits", "op": "", "content": "{\"splits\": [{\"dataset\": \"laion/dev_set_v2_a1_magicoder_20260805_125433\", \"config\": \"default\", \"split\": \"train\"}], \"pending\": [], \"failed\": []}", "coverage": ""}, {"obs_index": 1, "tool": "statistics", "op": "", "content": "{\"num_examples\": 32... | {"verdict": "pass", "refusal": false, "format": {"pass": true, "words": 28, "one_sentence": true}, "meta_language": {"pass": true, "matches": []}, "numbers": [{"text": "2", "normalized": "2", "badge": "ok", "sources": [{"obs_index": 0, "tool": "splits", "op": null, "partial": false}, {"obs_index": 1, "tool": "statistic... | 4 | pass | |||||
ca7726a7baa249b0820bb9046dd77743 | a | 0 | 2026-08-06T08:59:56.883170+00:00 | hysts-bot-data/daily-papers-embeddings | single | endpoint | LiquidAI/LFM2.5-2.6B | main | 0 | This dataset, hysts-bot-data/daily-papers-embeddings, contains daily academic papers (identified by arXiv IDs) with learned vector embeddings, enabling semantic search and retrieval across scholarly literature. | submit | [{"role": "user", "content": "Write ONE sentence (max 50 words, English) saying what this dataset is and what it is\nfor: domain, data source/provenance, method, languages, distinctive capability when stated.\nDo not restate license, popularity, or usage instructions. Do not invent anything \u2014 every claim\nmust com... | ["The user wants me to write a single sentence (max 50 words) describing what the dataset \"hysts-bot-data/daily-papers-embeddings\" is about - covering domain, data source/provenance, method, languages, and distinctive capabilities. I need to use the available tools to gather information about this dataset before prov... | [{"obs_index": 0, "tool": "splits", "op": "", "content": "{\"splits\": [{\"dataset\": \"hysts-bot-data/daily-papers-embeddings\", \"config\": \"default\", \"split\": \"train\"}], \"pending\": [], \"failed\": []}", "coverage": ""}, {"obs_index": 1, "tool": "statistics", "op": "", "content": "TOOL_ERROR: HTTPError: 404 C... | {"verdict": "pass", "refusal": false, "format": {"pass": true, "words": 23, "one_sentence": true}, "meta_language": {"pass": true, "matches": []}, "numbers": []} | 6 | pass | |||||
f894e2aeaabb40c8b1d681207365bf0d | a | 0 | 2026-08-05T20:50:07.509888+00:00 | HuggingFaceCode/stack-v3-train | single | endpoint | LiquidAI/LFM2.5-2.6B | main | 0 | The Stack v3 is the largest open source code dataset crawled directly from GitHub (August 2025), containing near-deduplicated, inline-file-containing code grouped by repository across 713 languages from 173M repositories—designed for pre-training code LLMs with full-repository context. | submit | [{"role": "user", "content": "Write ONE sentence (max 50 words, English) saying what this dataset is and what it is\nfor: domain, data source/provenance, method, languages, distinctive capability when stated.\nDo not restate license, popularity, or usage instructions. Do not invent anything \u2014 every claim\nmust com... | <img src="./assets/banner.png" width="100%" alt="The Stack v3 banner">
# 🥞 The Stack v3
- [What is it?](#what-is-it)
- [What is being released](#what-is-being-released)
- [How to download and use it](#how-to-download-and-use-it)
- [Dataset statistics](#dataset-statistics)
- [Dataset structure](#dataset-structure)
- ... | ["The user wants me to write ONE sentence (max 50 words) describing what this dataset is and what it is for, based on the provided card information. I need to:\n1. Ground my answer with tools - first check splits, then get statistics or other info\n2. Make sure every claim comes from the card or tool observations\n3. N... | [{"obs_index": 0, "tool": "splits", "op": "", "content": "{\"splits\": [{\"dataset\": \"HuggingFaceCode/stack-v3-train\", \"config\": \"default\", \"split\": \"train\"}], \"pending\": [], \"failed\": []}", "coverage": ""}, {"obs_index": 1, "tool": "statistics", "op": "", "content": "{\"num_examples\": 63312, \"statisti... | {"verdict": "pass", "refusal": false, "format": {"pass": true, "words": 36, "one_sentence": true}, "meta_language": {"pass": true, "matches": []}, "numbers": [{"text": "3", "normalized": "3", "badge": "ok", "sources": [{"obs_index": 0, "tool": "splits", "op": null, "partial": false}, {"obs_index": 1, "tool": "statistic... | 6 | good | 2026-08-05T20:50:09.821689+00:00 | pass | ||
5ccabc0736994dab97ed6488de6b11f6 | a | 0 | 2026-08-07T07:21:06.288661+00:00 | biglam/sloane-catalogues | single | endpoint | LiquidAI/LFM2.5-2.6B | main | 0 | The Sloane Manuscript Index Cards Dataset comprises digitized handwritten index cards from the British Library's Sloane Manuscript collection, providing archival catalog records with author names, manuscript titles, shelfmarks, and cross-references for OCR and VLM benchmarking. | submit | [{"role": "user", "content": "Write ONE sentence (max 50 words, English) saying what this dataset is and what it is\nfor: domain, data source/provenance, method, languages, distinctive capability when stated.\nDo not restate license, popularity, or usage instructions. Do not invent anything \u2014 every claim\nmust com... | # Sloane Manuscript Index Cards Dataset
## Dataset Description
This dataset contains digitized index cards from the British Library's Sloane Manuscript collection. The cards represent a catalog system used to index the Sloane manuscripts, one of the founding collections of the British Museum (now British Library).
#... | ["The user wants me to write ONE sentence (max 50 words) describing what this dataset is and its purpose, based on the provided card information. I need to:\n1. Ground my answer with tools - first check splits, then get statistics or sample rows to understand the dataset better\n2. Make sure every claim comes from the ... | [{"obs_index": 0, "tool": "splits", "op": "", "content": "{\"splits\": [{\"dataset\": \"biglam/sloane-catalogues\", \"config\": \"default\", \"split\": \"train\"}], \"pending\": [], \"failed\": []}", "coverage": ""}, {"obs_index": 1, "tool": "statistics", "op": "", "content": "TOOL_ERROR: HTTPError: 404 Client Error: N... | {"verdict": "pass", "refusal": false, "format": {"pass": true, "words": 35, "one_sentence": true}, "meta_language": {"pass": true, "matches": []}, "numbers": []} | 4 | pass | ||||
fd90f370a57442c094a1df66e30bfa91 | a | 0 | 2026-08-06T08:33:08.319870+00:00 | hysts-bot-data/daily-papers-embeddings | single | endpoint | LiquidAI/LFM2.5-2.6B | main | 0 | This dataset contains daily arXiv paper entries with learned vector embeddings, sourced from arXiv, enabling semantic similarity searches and retrieval across academic literature. | submit | [{"role": "user", "content": "Write ONE sentence (max 50 words, English) saying what this dataset is and what it is\nfor: domain, data source/provenance, method, languages, distinctive capability when stated.\nDo not restate license, popularity, or usage instructions. Do not invent anything \u2014 every claim\nmust com... | ["The user wants me to write a single sentence (max 50 words) describing what the dataset \"hysts-bot-data/daily-papers-embeddings\" is about - covering domain, data source/provenance, method, languages, and distinctive capabilities. I need to use the available tools to gather information about this dataset before prov... | [{"obs_index": 0, "tool": "splits", "op": "", "content": "{\"splits\": [{\"dataset\": \"hysts-bot-data/daily-papers-embeddings\", \"config\": \"default\", \"split\": \"train\"}], \"pending\": [], \"failed\": []}", "coverage": ""}, {"obs_index": 1, "tool": "statistics", "op": "", "content": "TOOL_ERROR: HTTPError: 404 C... | {"verdict": "pass", "refusal": false, "format": {"pass": true, "words": 23, "one_sentence": true}, "meta_language": {"pass": true, "matches": []}, "numbers": []} | 6 | good | 2026-08-06T08:36:54.047706+00:00 | pass | |||
91c20b049aaf43589492bba1083ccc72 | a | 0 | 2026-08-07T16:11:23.727938+00:00 | fzzhang/qwen3_8b_coding_depth4v3_probval_instill_n8_valredundancy5_round1 | single | endpoint | LiquidAI/LFM2.5-2.6B | main | 0 | This dataset (fzzhang/qwen3_8b_coding_depth4v3_probval_instill_n8_valredundancy5_round1) is a training split containing 31,218 examples, each with exactly two messages, focused on coding and probability tasks. | submit | [{"role": "user", "content": "Write ONE sentence (max 50 words, English) saying what this dataset is and what it is\nfor: domain, data source/provenance, method, languages, distinctive capability when stated.\nDo not restate license, popularity, or usage instructions. Do not invent anything \u2014 every claim\nmust com... | ["The user wants me to write a single sentence (max 50 words) describing what this dataset is - its domain, data source/provenance, method, languages, and distinctive capabilities. I need to use the available tools to gather information about this dataset before making my conclusion.\n\nLet me start by checking the spl... | [{"obs_index": 0, "tool": "splits", "op": "", "content": "{\"splits\": [{\"dataset\": \"fzzhang/qwen3_8b_coding_depth4v3_probval_instill_n8_valredundancy5_round1\", \"config\": \"default\", \"split\": \"train\"}], \"pending\": [], \"failed\": []}", "coverage": ""}, {"obs_index": 1, "tool": "statistics", "op": "", "cont... | {"verdict": "pass", "refusal": false, "format": {"pass": true, "words": 21, "one_sentence": true}, "meta_language": {"pass": true, "matches": []}, "numbers": [{"text": "3", "normalized": "3", "badge": "ok", "sources": [{"obs_index": 0, "tool": "splits", "op": null, "partial": false}, {"obs_index": 0, "tool": "splits", ... | 4 | good | 2026-08-07T16:12:02.949565+00:00 | pass | |||
4502cfc0cef34b77b50b465620f7ad5e | a | 0 | 2026-08-07T14:12:17.786834+00:00 | jablonkagroup/sfm-synthetic-data | single | endpoint | LiquidAI/LFM2.5-2.6B | main | 0 | The dataset jablonkagroup/sfm-synthetic-data is a collection of synthetic spectral data (likely from solid-state microscopy) designed for anomaly detection, organized into configs such as anomalous_gc, anomalous_hnmr, and others across five case splits, with each row containing spectral measurements and an anomaly flag... | submit | [{"role": "user", "content": "Write ONE sentence (max 50 words, English) saying what this dataset is and what it is\nfor: domain, data source/provenance, method, languages, distinctive capability when stated.\nDo not restate license, popularity, or usage instructions. Do not invent anything \u2014 every claim\nmust com... | ["The user wants me to write a single sentence (max 50 words) describing what the dataset \"jablonkagroup/sfm-synthetic-data\" is, covering:\n- Data source/provenance\n- Method\n- Languages\n- Distinctive capability\n\nI need to use the available tools to gather information about this dataset before writing the TLDR. L... | [{"obs_index": 0, "tool": "splits", "op": "", "content": "{\"splits\": [{\"dataset\": \"jablonkagroup/sfm-synthetic-data\", \"config\": \"anomalous_gc\", \"split\": \"case_1\"}, {\"dataset\": \"jablonkagroup/sfm-synthetic-data\", \"config\": \"anomalous_gc\", \"split\": \"case_2\"}, {\"dataset\": \"jablonkagroup/sfm-sy... | {"verdict": "pass", "refusal": false, "format": {"pass": true, "words": 46, "one_sentence": true}, "meta_language": {"pass": true, "matches": []}, "numbers": []} | 4 | pass | |||||
be11f36b6e4e47b79b9e1d9c6880ac10 | a | 0 | 2026-08-07T08:30:09.369764+00:00 | sahandjan/default-repo-dataset | single | endpoint | LiquidAI/LFM2.5-2.6B | main | 0 | This dataset, sourced from damafarm.ir, is a Persian-language collection of 20 dated records containing topics, probability scores, titles, and links, enabling analysis of time-stamped content. | submit | [{"role": "user", "content": "Write ONE sentence (max 50 words, English) saying what this dataset is and what it is\nfor: domain, data source/provenance, method, languages, distinctive capability when stated.\nDo not restate license, popularity, or usage instructions. Do not invent anything \u2014 every claim\nmust com... | # دامافارم Dataset
Source: [https://damafarm.ir](https://damafarm.ir) | ["The user wants me to write a single sentence (max 50 words) describing what this dataset is - its domain, data source/provenance, method, languages, and distinctive capabilities. I need to use the tools to gather information about this dataset before providing the TLDR.\n\nLet me start by checking the splits and basi... | [{"obs_index": 0, "tool": "splits", "op": "", "content": "{\"splits\": [{\"dataset\": \"sahandjan/default-repo-dataset\", \"config\": \"default\", \"split\": \"train\"}], \"pending\": [], \"failed\": []}", "coverage": ""}, {"obs_index": 1, "tool": "statistics", "op": "", "content": "{\"num_examples\": 20, \"statistics\... | {"verdict": "pass", "refusal": false, "format": {"pass": true, "words": 25, "one_sentence": true}, "meta_language": {"pass": true, "matches": []}, "numbers": [{"text": "20", "normalized": "20", "badge": "ok", "sources": [{"obs_index": 1, "tool": "statistics", "op": null, "partial": false}, {"obs_index": 1, "tool": "sta... | 5 | good | 2026-08-07T08:30:53.048459+00:00 | pass | ||
84137848b336469f8f6bbe506e524eaa | a | 0 | 2026-08-07T20:43:34.964989+00:00 | Hilti-Research/slam-challenge-2022 | single | endpoint | LiquidAI/LFM2.5-2.6B | main | 0 | The Hilti-Research/slam-challenge-2022 dataset is a SLAM (Simultaneous Localization and Mapping) benchmark containing 3,654 spatial coordinate records (x, y, z) designed for training and evaluating localization and mapping algorithms. | submit | [{"role": "user", "content": "Write ONE sentence (max 50 words, English) saying what this dataset is and what it is\nfor: domain, data source/provenance, method, languages, distinctive capability when stated.\nDo not restate license, popularity, or usage instructions. Do not invent anything \u2014 every claim\nmust com... | ["The user wants me to write a single sentence (max 50 words) describing what the dataset \"Hilti-Research/slam-challenge-2022\" is about - covering domain, data source/provenance, method, languages, and distinctive capabilities. I need to use the available tools to gather information about this dataset before providin... | [{"obs_index": 0, "tool": "splits", "op": "", "content": "{\"splits\": [{\"dataset\": \"Hilti-Research/slam-challenge-2022\", \"config\": \"default\", \"split\": \"train\"}], \"pending\": [], \"failed\": []}", "coverage": ""}, {"obs_index": 1, "tool": "statistics", "op": "", "content": "TOOL_ERROR: HTTPError: 404 Clien... | {"verdict": "pass", "refusal": false, "format": {"pass": true, "words": 28, "one_sentence": true}, "meta_language": {"pass": true, "matches": []}, "numbers": [{"text": "2022", "normalized": "2022", "badge": "ok", "sources": [{"obs_index": 0, "tool": "splits", "op": null, "partial": false}, {"obs_index": 1, "tool": "sta... | 7 | pass |
End of preview. Expand in Data Studio
No dataset card yet
- Downloads last month
- 383