Datasets:

Catherine1213
/

VerifierBench

language:
  - en
  - zh
license: other
size_categories:
  - 1K<n<10K
task_categories:
  - text-generation
license_name: apache-2.0-and-upstream-licenses
license_link: LICENSE.md
tags:
  - verifier-generation
  - code
  - evaluation
  - benchmark
configs:
  - config_name: tasks
    data_files:
      - split: full
        path: data/tasks.parquet
    default: true
  - config_name: candidates
    data_files:
      - split: full
        path: data/candidates.parquet

VerifierBench

Paper · Project page · Code

VerifierBench provides the paper’s frozen 222 tasks and 4,993 constructed candidates with PASS/FAIL labels. AppWorld content requires local reconstruction; source terms are in LICENSE.md.

from datasets import load_dataset

tasks = load_dataset("Catherine1213/VerifierBench", "tasks", split="full")
candidates = load_dataset("Catherine1213/VerifierBench", "candidates", split="full")