AI Dataset2026
Benchmark dataset for evaluating long-horizon coding agents and software-engineering tasks, containing English code and tabular metadata in Parquet format; small scale (<1K examples) for fast prototyping and evaluation.
AI Dataset2026
De-identified longitudinal multimodal CT dataset for multicancer screening that pairs ~24k CT volumes with radiology reports and voxel-wise tumor annotations across 13 cancer types. Designed for longitudinal disease modeling, detection/segmentation and vision–language research; CC BY‑NC‑ND 4.0 for non-commercial use.
AI Audio2026
Local English text-to-waveform TTS producing a single fixed synthetic voice in a deployable package below 10M parameters. Offers deterministic seeds, punctuation-aware long-text chunking, CPU/CUDA and ONNX runtime options, measured evaluations and a compact FP32 footprint; English-only, one voice.
AI Dataset2026
Provides runnable 3D programs with renders, descriptions, and metadata for training and evaluating models that generate 3D code from text or images. Dataset access is restricted to non-commercial research.
AI Dataset2026
Provides 2,165 historical natural-history page scans paired with ~99.95% expert transcriptions and pixel-aligned PAGE XML layout ground truth for OCR and layout evaluation; multilingual (EN/FR/DE/LA), CC-BY 3.0.
AI Dataset2026
Tests whether language models can solve original, culturally grounded riddles—not just translate them. It spans 62 communities and combines model answers with native-speaker evaluation; use is restricted to evaluation.
Proposes Riemannian Isometric Policy Optimization (RIPO) to fix exploration collapse in PPO-style RL for LLMs by aligning policy updates with the policy manifold's Riemannian geometry, improving exploration–exploitation balance and optimization stability across competition benchmarks.
Builds structured knowledge graphs for retrieval-augmented generation via a multi-step GraphRAG pipeline that separates extraction from consolidation. Key features include typed two-stage extraction, DBSCAN-backed deduplication, LLM summarization, Leiden community detection, and a compact 7B extractor model (Meno-Lite-0.1).
AI Model2026
Generates 2048-d multilingual text embeddings for retrieval and semantic search, suited for RAG and dense retrieval. Pruned and distilled from the Ministral-3 family into a ~1.14B BF16 model, supports long contexts (up to 32,768 tokens) and optimized for NVIDIA GPU inference.
AI Dataset2026
Evaluates retrievers and search agents on synthetic multi-hop questions that require assembling a complete set of supporting evidence. Provides English and Russian variants (395 questions each), a fixed dense index embedded with Qwen3-Embedding-8B, and BrowseComp-Plus evaluation integrations.
AI Video Papers2026
Evaluates whether video models reason according to physical laws by treating generated videos as visible reasoning traces and using a three-stage Perception–Formulation–Deduction protocol. Includes Orchard (400 mechanics videos), chain-of-frames prompting on annotated first frames, and a hybrid MLLM-plus-objective scoring suite for stage-resolved diagnostics.
Quantifies active visual observation in multimodal LLMs with ActiveVision, a 17-task benchmark that forces repeated perception rather than one-shot description. Finds frontier MLLMs fail badly (top model 10.6% vs humans 96.1%) and that model-generated vision code does not close the gap.