Datasets:
Tasks:
Text Classification
Formats:
parquet
Languages:
English
Size:
1K - 10K
Tags:
structured-decisions
calibration
probabilistic-classification
system-one
workflow-evaluation
Synthetic
License:
YAML Metadata Error:Invalid content in eval.yaml.
Check out the documentation for more information.
Show details
✖ `evaluation_framework` must be omitted or one of: exgentic, inspect-ai, math-arena, mteb, olmocr-bench, harbor, ifstruct, pier, redline-bench, archipelago, benchflow, apex-evals, screenspot-pro, swe-bench, swe-bench-pro, nemo-evaluator, yc-bench, open-asr-leaderboard, mdpbench, real5-omnidocbench, parsebench, video-mme-v2, claw-eval, researchclawbench, pbench, wildclawbench, wbench, nanofold, mmmu, extractbench
→ at evaluation_framework
| name: Typed Decisions | |
| description: > | |
| A benchmark for typed probabilistic decisions. A model gets one piece of | |
| unstructured state and answers five typed questions about it at once, and | |
| every answer is a probability distribution, not a single label. | |
| evaluation_framework: none | |
| tasks: | |
| - id: accuracy | |
| config: all | |
| split: test | |
| - id: kl_from_gold | |
| config: all | |
| split: test | |
| lower_is_better: true | |
| - id: brier | |
| config: all | |
| split: test | |
| lower_is_better: true | |
| - id: ece | |
| config: all | |
| split: test | |
| lower_is_better: true | |