MT-Mind2Web

Multi-turn web-navigation conversations constructed from Mind2Web interactions, with dialogue turns, HTML snapshots, element candidates and click/type/select action sequences.

AI-assisted editorial review against a primary source. This is a catalog annotation, not an independent reproduction or endorsement.

What is this resource?

Resource type

Datasets

Environment

web

Topics

web-navigation, imitation-learning, memory

Record publication date

Not recorded

Evidence retrieved

2026-10-17T12:34:10Z

Resource ID

url:bdfb50330deb97893ce4

What licenses are recorded?

Code, training data, weights, and documents can have different terms. A blank declaration means unknown, not unrestricted use.

Code

Not recorded

Data

cc-by-4.0 (Publisher dataset-card declaration for the multi-turn annotation release.)

Document

Not recorded

Weights

Not recorded

License review notes

• Data: Publisher dataset-card declaration for the multi-turn annotation release. Source 1

Dataset profile

Supervision

multi-turn web instruction following, element/action prediction

Modalities

dialogue text, HTML, element candidates, actions

Formats

JSON, auxiliary PKL

Splits

train, test_task, test_website, test_subdomain

Scale and counting units

• 600 conversations — train source
• 34 conversations — test-task source
• 42 conversations — test-website source
• 44 conversations — test-subdomain source
• 2,896 dialogue turns — train source
• 191 dialogue turns — test-task source
• 218 dialogue turns — test-website source
• 216 dialogue turns — test-subdomain source

Conversation, dialogue-turn and action-step units differ. The released file manifest is inspected; underlying examples were not downloaded.

Compare source-reviewed profiles

Recorded specifications

availability

Ungated JSON conversation shards and ranking-file manifest.

release revision

5ae9a3bedf431f54c2218fbb8ee474e5efbda3a8

Explore related topics

Scope and limitations

The ungated Hub manifest contains train, test-task, test-website and test-subdomain JSON shards plus a candidate-ranking pickle file. Counts distinguish conversation sessions from dialogue turns; neither is a screenshot or action-step count. No pickle or sample data was loaded. The card declares CC-BY-4.0 and identifies the original Mind2Web interactions as its source.

Related resources

• documented by: On the Multi-turn Instruction Following for Conversational Web Agents

Connected assets

These links come from reviewed resource records. A documentation link does not imply a training-data or evaluation dependency.

• uses: Self-MAP · Frameworks · relationship evidence

Primary evidence and verification

Source retrieval is separate from experiment reproduction. Use the original publication or publisher page when citing claims about the resource.

1. editorial primary review

   Retrieved 2026-10-17T12:34:10Z. Supports: name, kind, summary, environments, topics, scope, notes, details.dataset_profile, licenses.data.

   Content fingerprint

   8685e0a26a06b3339b5c078632a7b0ade1ead40e6b586ef1ff3676414699a4bc
2. editorial primary review

   Retrieved 2026-10-17T12:34:10Z. Supports: identity, release_status, details.release_revision.

   Content fingerprint

   4455e6daf1151694c03b2a1ebd4a7af40b5b079b0568706a1783a0222383bb1d
3. editorial primary review

   Retrieved 2026-10-16T06:36:58Z. Supports: relations.

   Content fingerprint

   c87e97eb397da04806d53145cf9c13be739fc657bb54dca9318c8c4ed2a36249

Suggest a correction

Open a source-backed correction. Include this record ID and the original source. Read our editorial policy and coverage boundaries.