{ "@context": { "ml": "http://mlcommons.org/croissant/", "schema": "https://schema.org/", "sc": "https://schema.org/", "codesense": "https://codesense-bench.github.io/schema/" }, "@type": "schema:Dataset", "schema:name": "CodeSense-Bench-loop_body_dataset_python", "schema:description": "Dataset for In-Loop value reasoning of LLMs", "schema:license": "CC-BY-4.0", "schema:url": "https://huggingface.co/datasets/codesense-bench/codesense/blob/main/Task%20Specific%20Datasets/loop_body_dataset_python.jsonl", "schema:creator": { "@type": "schema:Organization", "schema:name": "CodeSense Research Team" }, "schema:distribution": [ { "@type": "schema:FileObject", "schema:contentUrl": "https://huggingface.co/datasets/codesense-bench/codesense/resolve/main/Task%20Specific%20Datasets/loop_body_dataset_python.jsonl", "schema:encodingFormat": "application/jsonl", "schema:name": "loop_body_dataset_python", "schema:sha256": "4eca6a5382cecc7b4a54fc269093de0ed7f06c1ba89226f3aff3353c8ee84b29" } ], "recordSet": [ { "name": "loop_body_dataset_python", "description": "A collection of Python code examples designed to evaluate a model’s ability to reason about variable values inside loop execution. ", "field": [ { "name": "idx", "dataType": "sc:Integer", "description": "Unique identifier for each record in the dataset." }, { "name": "scratchpad_format", "dataType": "sc:Text", "description": "code in scratchpad format, the variable states defined in comments." }, { "name": "loop_code", "dataType": "sc:Text", "description": "code used for LLM prompting." }, { "name": "question", "dataType": "sc:Text", "description": "Query regarding a specific line in-loop in the code." }, { "name": "answer", "dataType": "sc:Text", "description": "Ground-Truth, contains the exact value of the variable in-loop execution." }, { "name": "variable_assignment", "dataType": "sc:Text", "description": "selected variables value as assignment." } ] } ] }