# hypobench-few_shot-gpt5420260305-full-20260415_015532-predictions

Benchmark: hypobench
Generated: 2026-04-15T02:13:41.456606

Per-test-row predictions from practical utility evaluation.
Parent dataset: [hypobench-few_shot-gpt5420260305-full-20260415_015532](https://huggingface.co/datasets/strategy-scope/hypobench-few_shot-gpt5420260305-full-20260415_015532)

## Columns
- **task**: task name
- **index**: test row index
- **true_label**: ground-truth label
- **pred_label**: model prediction
- **correct**: 1 if prediction matches ground truth, 0 otherwise
- **raw_response**: full LLM response