218 evaluation samples support the evaluation of models on Sentence Stress Reasoning (SSR) and Sentence Stress Detection (SSD) tasks. The dataset is associated with the paper 'StressTest: Can YOUR Speech LM Handle the Stress?' and was created by author slprl. It was last updated on April 8, 2026.
Use Cases
- Benchmarking model performance on Sentence Stress Reasoning (SSR) tasks.
- Evaluating model accuracy on Sentence Stress Detection (SSD) tasks.
- Training or fine-tuning speech language models to handle prosodic variation.
Strengths
- Dataset size is explicitly stated as 218 evaluation samples.
- Specifically designed for two defined tasks: Sentence Stress Reasoning (SSR) and Sentence Stress Detection (SSD).
Limitations
- Column-level documentation is absent; field semantics must be inferred after download.
- Row count beyond the 218 test samples is unknown, which may limit suitability assessment.
Provenance
- Source
- slprl
- Freshness
- Last updated 2026-04-08 12:20:06.