Benchmarks

ScreenSpot-v2 instruction variants

Instruction variants over ScreenSpot-v2 screenshots that distinguish original commands, explicit actions, target descriptions and impossible operations.

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

Original source ↗Read as MarkdownRecord JSON

What is this resource?

Resource type
Benchmarks
Environment
cross-platform
Topics
grounding, evaluation, robustness
Record publication date
Not recorded
Evidence retrieved
2026-10-17T04:07:19Z
Resource ID
url:17729e1c532ba7043404

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
Not recorded
Document
Not recorded
Weights
Not recorded

Evaluation protocol

Focus
instruction-form robustness, negative grounding requests
Setting
offline
Observations
screenshots, instructions
Actions / outputs
target localization, identification of impossible requests
Evaluation
Publisher supplies original, action, description and negative instruction types; no common end-to-end success metric is asserted.

Scale and counting units

Reuses ScreenSpot-v2 imagery. Image-file count is not treated as a count of independent screenshots or tasks.

Compare source-reviewed profiles

Recorded specifications

disabled
false
download url
https://huggingface.co/datasets/likaixin/ScreenSpot-v2-variants/tree/main
gated
false
hub id
likaixin/ScreenSpot-v2-variants

Explore related topics

GUI groundingEvaluationRobustness

Scope and limitations

Public image files and desktop/mobile/web annotation JSON files are listed. This is a changed instruction set over an existing screenshot collection, not an independent new screenshot corpus. No data license or precise example count is declared in the card; no paper relationship is inferred from its generic project/PDF links.

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 2026-10-17T04:07:19Z. Supports: name, kind, summary, topics, scope, relations, details.

    Content fingerprint7fb2bc9499d8a2b37b663c116af49d11bba493e6767d76a5305fad93257bebf2
  2. editorial primary review

    Retrieved 2026-10-17T04:07:19Z. Supports: identity, release_status, details, environments.

    Content fingerprintdea98efa75e063eee2d6bfb5d46da94b41df3d779e2fb0bdaea7866534b5c469

Suggest a correction

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