Datasets:
|
Download task_contract.md from QLU-LLM/MAGAR-Bench: direct link, hf CLI and curl.
- Browser
- Download file 2.44 kB
-
https://huggingface.co/datasets/QLU-LLM/MAGAR-Bench/resolve/main/task_contract.md
- Command line
-
hf download hf://datasets/QLU-LLM/MAGAR-Bench/task_contract.md
-
curl -L -o task_contract.md https://huggingface.co/datasets/QLU-LLM/MAGAR-Bench/resolve/main/task_contract.md
2.44 kB
| # MAGAR-Bench task contract | |
| The public input format is shared across all three task families. Each record | |
| contains a `messages` array, an `images` array, and metadata. This Test-only export already contains user messages only. The model must emit one JSON object only, | |
| without Markdown, explanations, wrapper objects, or extra fields. | |
| ## Assembly Decision | |
| Required output fields: | |
| ```json | |
| { | |
| "operation": "place", | |
| "involved_part_ids": ["<part id>"], | |
| "next_subgoal": "place base part" | |
| } | |
| ``` | |
| `operation` is one of `align`, `execute_subtask`, `inspect`, `lock`, `place`, or | |
| `recover`. `next_subgoal` is one of the canonical values listed in | |
| `schema.json`. High-level task IDs are excluded from this release metadata and from the | |
| model output contract. Where a legacy prompt first asks for that field, the | |
| later explicit vocabulary clause excluding it is authoritative. | |
| ## Atomic Action | |
| Required output is one complete native GUI primitive. The action vocabulary is | |
| `click`, `drag`, `hotkey`, `keypress`, and `type_text`. The action type and its | |
| parameters must stay together; future actions and action sequences are not | |
| allowed. Coordinates are screen pixels when the action requires coordinates. | |
| Examples of the native shapes are: | |
| ```json | |
| {"type": "click", "button": "left", "screen_point_px": [320, 240]} | |
| ``` | |
| ```json | |
| {"type": "keypress", "key": "R"} | |
| ``` | |
| ## Minimal Rework | |
| Required output fields: | |
| ```json | |
| { | |
| "objective": "minimum_distinct_moved_parts", | |
| "selected_route_id": "<route id>", | |
| "moved_part_ids": ["<part id>"], | |
| "min_moved_part_count": 1, | |
| "action_ids": ["<action id>"], | |
| "full_shortest_path_claim": false | |
| } | |
| ``` | |
| The moved-part count must equal the cardinality of `moved_part_ids`. The | |
| reference is the minimum distinct-part solution inside the validated recovery | |
| candidate set for that state. This benchmark does not claim unrestricted | |
| global optimality over every possible GUI action sequence. Route and action | |
| identifiers are auxiliary recovery information; matching them does not execute | |
| the recovery. | |
| ## Split and privacy rules | |
| The 20 Train families and six Test families are disjoint. `test.jsonl` | |
| contains no assistant answers or Test targets. The locked Test gold must stay in | |
| a separate gated evaluator or private artifact. | |
| Only the six Test families are distributed here. No Train records or reference answers are included. | |