A dataset that quantitatively compares the intelligence and inference differences between speech‑input and text‑input in voice language models by applying TTS to samples from three benchmarks (Elyza‑tasks‑100, M‑IFEval, JamC‑QA), producing four subsets—Elyza, Spoken‑Elyza, M‑IFEval, and JamC‑QA—with 1,452 usable audio examples under a CC BY‑SA 4.0 license for non‑commercial use.
Description in other languages
Statistics
Statistics as of Nov 16, 2026
Hugging Face downloads (last 30 days)
161
Hugging Face likes
9
Listed
Apr 6, 2026
Downloads (30-day window) over time
Show as a table
| Month | DL (30d) |
|---|---|
| Dec 2025 | — |
| Jan 2026 | — |
| Feb 2026 | — |
| Mar 2026 | — |
| Apr 2026 | 187 |
| May 2026 | 306 |
| Jun 2026 | 813 |
| Jul 2026 | 529 |
| Aug 2026 | 268 |
| Sep 2026 | 102 |
| Oct 2026 | 188 |
| Nov 2026 | 161 |
Downloads over the 30 days before the last snapshot of each month.
Popular in the same category
- JGLUEshunk0313kUpdated JGLUE dataset card and loading script for a Japanese NLP benchmark (created by Yahoo Japan and Waseda University) that covers text classification (MARC‑ja, JCoLA), sentence‑pair classification (JNLI), and QA (JSQuAD, JCommonsenseQA), with releases linked on GitHub and Hugging Face.
- JamC-QAsbintuitions2kJamC‑QA is a bilingual benchmark of multiple‑choice questions spanning eight Japanese‑culture and knowledge categories, with leaderboard metrics comparing state‑of‑the‑art models.
- JMMMU-ProJMMMU1kJMMMU‑Pro is an image‑based Japanese multimodal benchmark created by generating visual questions with a realistic model and human verification, revealing that current open‑source LMMs perform poorly on it and offering a cost‑effective method for future VQA benchmark development.
- JMedBenchColdog23331kJMedBench is a Japanese biomedical LLM benchmark comprising 20 datasets across five tasks (MCQA, NER, STS, etc.) sourced from MedMCQA, PubMedQA, MMLU, and others, each with its own license, and includes a note that translations may contain biases requiring human review.
- JAMMEvalllm-jp830JAMMEval is a distilled benchmark of seven Japanese VQA datasets, refined with two rounds of human annotation to remove ambiguities and non‑visual questions, delivering reliable evaluation of vision‑language models on multimodal Japanese tasks.
- llm-jp-evalllm-book666Dataset card for the ja‑vicuna‑qa‑benchmark used in the book “Introduction to Large‑Scale LLM II” and created by llm‑jp‑eval for cross‑dataset Japanese LLM evaluation (Apache 2.0).
Resources in similar research fields
- JA-Multi-Image-VQASakanaAI311JA‑Multi‑Image‑VQA is a dataset of 39 images and 55 Japanese question‑answer pairs for evaluating multi‑image VQA, licensed under Apache 2.0 (except the images).Shared: Annotation and Dataset Development
- JDocQA-Reasoningricoh-ai269A Japanese benchmark, JDocQA Reasoning Bench, released by Ricoh for the GENIAC project, evaluates LMMs on 1,286 reasoning questions over complex charts, tables, and diagrams using automated LLM-as‑Judge scoring.Shared: Annotation and Dataset Development
- ELYZA-tasks-100elyza2kA 100‑sample Japanese instruction‑tuning evaluation dataset of annotated tasks—ranging from summarization correction and math reasoning to translation, creative generation, and user‑intent understanding—designed for manual or automatic 5‑point rating of fine‑tuned models.Shared: Annotation and Dataset Development
- embhpprc1kA catalog of Japanese and multilingual QA, NLI and paraphrase datasets, detailing each dataset’s retrieval or QA tasks and its license (Apache 2.0, CC‑BY‑SA/CC‑BY, MIT, etc.).Shared: Annotation and Dataset Development
- voicevox-voice-corpusayousanz885VOICEVOX‑generated synthetic voice dataset comprising 445,793 .wav files (totaling 577 h 51 m 23 s) built from the ITA, つくよみちゃん, and ROHAN corpora.Shared: Annotation and Dataset Development
lawqa_jpdigital-go-jp279Multiple-choice QA dataset on Japanese regulationsGitHubShared: Annotation and Dataset Development