Datasets:
task_name stringlengths 10 28 | status stringclasses 3
values | passed bool 2
classes | messages listlengths 17 372 | original_message_count int32 17 372 | readable_message_count int32 17 372 | tool_call_count int32 10 292 | initial_run_time stringdate 2026-08-02 17:54:21 2026-08-02 17:58:27 | completion_time stringdate 2026-08-02 17:55:16 2026-08-02 18:03:44 | model stringclasses 1
value | served_model_name stringclasses 1
value | slurm_job_id stringclasses 1
value | toolathlon_job_id stringclasses 1
value | run_context_length_http_400_count int32 114 114 | infrastructure_compromised bool 1
class |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
ab-testing | success | true | [
{
"role": "user",
"content": "The A/B test for our new homepage has concluded, and the raw clickstream data has been stored in the `ab_testing` dataset in BigQuery. Analyze the data and fill the `record.csv`, determine which version ('A' or 'B') has the highest overall conversion rate, i.e., the overall con... | 66 | 66 | 39 | 2026-08-02 17:55:46 | 2026-08-02 18:03:23 | Qwen/Qwen3.5-4B | qwen3.5-4b | 95538 | job_b51949fc035f | 114 | true |
arrange-workspace | success | false | [
{
"role": "user",
"content": "Please help me organize the files in my workspace. The file structure should be as shown below. \n\n├── School\n│ ├── Courses_Materials\n│ ├── Graduation_Projects\n│ ├── Applications_Materials\n│ └── Language_Exam_Preparation\n├── Work\n│ ├── Projects\n│ ├── Softwar... | 76 | 76 | 63 | 2026-08-02 17:54:21 | 2026-08-02 17:56:11 | Qwen/Qwen3.5-4B | qwen3.5-4b | 95538 | job_b51949fc035f | 114 | true |
excel-data-transformation | success | false | [{"role":"user","content":"Please convert the file `Household_Appliances.xlsx` from its current two-(...TRUNCATED) | 165 | 165 | 88 | 2026-08-02 17:56:12 | 2026-08-02 18:03:36 | Qwen/Qwen3.5-4B | qwen3.5-4b | 95538 | job_b51949fc035f | 114 | true |
experiments-recordings | failed | false | [{"role":"user","content":"In the table of the Notion page `mcp_experiments_recordings`, based on th(...TRUNCATED) | 174 | 174 | 94 | 2026-08-02 17:56:40 | 2026-08-02 18:00:54 | Qwen/Qwen3.5-4B | qwen3.5-4b | 95538 | job_b51949fc035f | 114 | true |
find-alita-paper | success | false | [{"role":"user","content":"I'm looking for a paper on arxiv related to agentic reasoning. I only rem(...TRUNCATED) | 24 | 24 | 11 | 2026-08-02 17:56:36 | 2026-08-02 17:57:25 | Qwen/Qwen3.5-4B | qwen3.5-4b | 95538 | job_b51949fc035f | 114 | true |
flagged-transactions | success | true | [{"role":"user","content":"Perform anomaly detection on high-net-worth clients’ transactions: Extr(...TRUNCATED) | 44 | 44 | 22 | 2026-08-02 17:54:29 | 2026-08-02 17:59:14 | Qwen/Qwen3.5-4B | qwen3.5-4b | 95538 | job_b51949fc035f | 114 | true |
git-milestone | success | true | [{"role":"user","content":"In recent days, the total number of repos on GitHub has reached 1B, which(...TRUNCATED) | 17 | 17 | 10 | 2026-08-02 17:54:21 | 2026-08-02 17:55:16 | Qwen/Qwen3.5-4B | qwen3.5-4b | 95538 | job_b51949fc035f | 114 | true |
invoice-org | success | false | [{"role":"user","content":"Please help me read the files in my personal workspace, find all of my in(...TRUNCATED) | 24 | 24 | 16 | 2026-08-02 17:54:21 | 2026-08-02 17:56:03 | Qwen/Qwen3.5-4B | qwen3.5-4b | 95538 | job_b51949fc035f | 114 | true |
k8s-deployment-cleanup | success | true | [{"role":"user","content":"Find all Deployments in namespaces with names starting with 'dev-' that a(...TRUNCATED) | 84 | 84 | 58 | 2026-08-02 17:56:20 | 2026-08-02 17:58:35 | Qwen/Qwen3.5-4B | qwen3.5-4b | 95538 | job_b51949fc035f | 114 | true |
notion-find-job | failed | false | [{"role":"user","content":"You are a professional job search assistant. You need to manage a Notion (...TRUNCATED) | 137 | 137 | 61 | 2026-08-02 17:54:31 | 2026-08-02 17:59:38 | Qwen/Qwen3.5-4B | qwen3.5-4b | 95538 | job_b51949fc035f | 114 | true |
End of preview. Expand in Data Studio
Toolathlon Qwen3.5-4B Pass@1 Pilot
Readable trajectories from a 15-task Toolathlon-Verified Pass@1 pilot using
Qwen/Qwen3.5-4B in thinking mode. The model was served on 6 B200 GPUs with
data parallelism; Toolathlon's Docker environments ran on its remote public
evaluation service.
Contents
- Repository:
violetxi/toolathlon-qwen35-4b-base-pass1-pilot15 - Rows: 15 (one row per task)
- Reported result: 5/15 (33.33%)
messages: a Viewer-friendly list containing onlyuserandassistantroles. Tool calls are rendered inside assistant messages; tool results are rendered as labeled user messages so the full interaction remains readable.- Per-row task status, pass/fail, timing, and tool-call counts are included.
Important infrastructure caveat
This pilot is not a clean performance measurement. vLLM used a 65,536-token context window while each request reserved up to 32,768 output tokens. The run recorded 114 HTTP 400 context-length errors once prompts exceeded the remaining input budget. The trajectories are useful for qualitative inspection, but the reported 5/15 Pass@1 should not be used as a canonical model score.
Transformation and safety
The source trajectories mixed user, assistant, and tool records. This
dataset converts them to a two-role readable representation and scans/redacts
credential-shaped strings. Task workspaces, authentication files, and raw
environment archives are not uploaded.
- Downloads last month
- 48