All datasets / zetomatoz

zetomatoz/guidellm-agentic-coding-trajectories

98 downloads in 30 days5 likesLast modified 29 Sep 2026View on Hugging Face

Unclear

The checks could not identify a recognised licence ID in the dataset's metadata.

Declared licence: other, from the tagEvidence: README.md at 94fe7ff
Licence name on the card: derivative-multi-source, link: LICENSEEvidence: README.md at 94fe7ff
Licence files: LICENSE, detected as cc-by-4.0Evidence: LICENSE at 94fe7ff
Declared sources: None listed in the card metadata
Access gated on Hugging Face: No

Each cell links to what was checked: a file, a file list or a Hub record.

Tasks

text-generation

Citation text from the card

This is an adapted, sampled benchmark derived from [Thoughtworks/agentic-coding-trajectories](https://huggingface.co/datasets/thoughtworks/agentic-coding-trajectories) (revision `cef72d1f4d0caabf85937adf8337a14b7522c782`), itself a derivative of three upstream trajectory datasets. Changes here: sample and deduplicate sessions, count output targets with a Qwen tokenizer, and select the original sample using an estimated context budget. This revision recovers native tool-call metadata and recorded results for the sampled sessions, packs conversations into GuideLLM graphs, substitutes one mock function for native tools, and moves the existing fixed output targets to text continuations. Text-command sources retain their original flattened prompts. Context estimates are recalculated for the expanded workload. These files include original trajectory text. Credit Thoughtworks and the upstream dataset creators when reusing or sharing them.

`license: other` and `license_name: derivative-multi-source` describe the mixed provenance; they do **not** grant a single new license for all rows. See [LICENSE](LICENSE) for source links and terms. Every row includes `source_dataset` and `source_id` so its upstream source can be identified. The 8, 24, and 48 turn configurations contain 90, 450, and 90 rows respectively; source records can occur in more than one configuration.

| `source_dataset` | Upstream trajectories | Published dataset license | Rows in 8 / 24 / 48 turn configs |
| --- | ---

All datasets marked Unclear | JSON

The same facts for your agent

A free MCP server at mcp.datalicense.uk/mcp. Connect it from an MCP client such as Claude or Cursor.

dataset_licence({ "id": "zetomatoz/guidellm-agentic-coding-trajectories" })
signal: "unclear"
stated_licence: "other"
sources: 0