# 📊 Eval / Observability / Safety

- [LiquidAI/ifstruct-v1.0](https://cenxialiu7-cloud.github.io/ai-oss-daily/en/p/dataset-liquidai-ifstruct-v1-0/) — · benchmark:official, benchmark:eval-yaml, task_categories:text-generation
- [redmadrobot-rnd/pii_benchmark](https://cenxialiu7-cloud.github.io/ai-oss-daily/en/p/dataset-redmadrobot-rnd-pii-benchmark/) — · task_categories:token-classification, language:ru, license:mit
- [Rapidata/svg-benchmark](https://cenxialiu7-cloud.github.io/ai-oss-daily/en/p/dataset-rapidata-svg-benchmark/) — · task_categories:text-to-image, task_categories:image-classification, task_categories:reinforcement-learning
- [sbintuitions/joyo-kanji-yomi-benchmark](https://cenxialiu7-cloud.github.io/ai-oss-daily/en/p/dataset-sbintuitions-joyo-kanji-yomi-benchmark/) — · task_categories:text-to-speech, language:ja, license:mit
- [treble-technologies/ffasr](https://cenxialiu7-cloud.github.io/ai-oss-daily/en/p/space-treble-technologies-ffasr/) — · gradio, leaderboard, asr
- [nvidia/Cosmos-HumanEval-v1](https://cenxialiu7-cloud.github.io/ai-oss-daily/en/p/dataset-nvidia-cosmos-humaneval-v1/) — · region:us
- [thenicolas1894/awesome-claude-fable-5-prompt-vault](https://cenxialiu7-cloud.github.io/ai-oss-daily/en/p/thenicolas1894-awesome-claude-fable-5-prompt-vault/) — Ultimate Claude Fable 5 Guide 2026: Use Cases, Integrations & Benchmarks
- [ScaleAI/SWE-bench_Pro](https://cenxialiu7-cloud.github.io/ai-oss-daily/en/p/dataset-scaleai-swe-bench-pro/) — · benchmark:official, benchmark:eval-yaml, size_categories:n<1K
- [cais/hle](https://cenxialiu7-cloud.github.io/ai-oss-daily/en/p/dataset-cais-hle/) — · benchmark:official, license:mit, size_categories:1K<n<10K
- [gaia-benchmark/GAIA](https://cenxialiu7-cloud.github.io/ai-oss-daily/en/p/dataset-gaia-benchmark-gaia/) — · language:en, size_categories:n<1K, format:parquet
- [Idavidrein/gpqa](https://cenxialiu7-cloud.github.io/ai-oss-daily/en/p/dataset-idavidrein-gpqa/) — · benchmark:official, benchmark:eval-yaml, task_categories:question-answering
- [MadsLorentzen/ai-job-search](https://cenxialiu7-cloud.github.io/ai-oss-daily/en/p/madslorentzen-ai-job-search/) — The job search that runs on your machine. AI job application framework built on Claude Code: evaluate posting…
- [ibm-research/ScarfBench](https://cenxialiu7-cloud.github.io/ai-oss-daily/en/p/dataset-ibm-research-scarfbench/) — · task_categories:text-generation, arxiv:2605.06754, region:us
- [MME-Benchmarks/Video-MME-v2](https://cenxialiu7-cloud.github.io/ai-oss-daily/en/p/mme-benchmarks-video-mme-v2/) — Video-MME-v2: Towards the Next Stage in Benchmarks for Comprehensive Video Understanding