AI Dataset2026
Aggregated, screened corpus of 55,050 normalized Indian public-information text bodies and 65,209 source records for retrieval and question-answering. Exports include deduplicated CSV/Parquet with provenance, topic labels, extraction quality flags and a private SQLite backup.
AI Dataset2026
Provides a monthly Parquet snapshot of ~5.6 billion public TikTok videos (2014–Oct 2026), including captions, hashtags, sounds, engagement metrics and TikTok Shop links. Designed for large-scale querying (DuckDB/Pandas/Polars); licensed CC BY-NC 4.0 for research and personal use.
AI Dataset2026
Provides 16 weeks of anonymized production agent-session traces (12,002 sessions, ~1.19M LLM requests, ~1.21M tool calls, 209B input tokens) released as block-level prefix IDs plus flattened Parquet tables for KV-cache, scheduling and serving-system research.
AI Dataset2026
Evaluation dataset for comparing eight text-to-image models using 8,000 generated images with source prompts and per-image scores for aesthetic quality, emotional resonance, and content integrity. Includes model labels, shared prompts, GPT-5.6 Sol automated scores, embedded images in Parquet, and an Apache-2.0 license.
AI Dataset2026
Provides 369 Harbor sandbox tasks ported from OpenAI's openai/math: each task is a Lean theorem with missing `sorry` proofs that an agent must complete, graded by a strict Comparator exact-match verifier. Includes task definitions, generator, and manifest for RL/code-agent evaluation.