Tag

MathNet v0 — Olympiad Math Reasoning & Retrieval

Provides a 30K+ problem multimodal, multilingual dataset of Olympiad-level math problems with expert solutions and a math-aware retrieval benchmark—includes images, hierarchical topics, provenance from official booklets, and LLM-assisted metadata (v0, CC BY 4.0).

DeskcommCRM

Self-hosted sales CRM that runs native AI agents (RAG per tenant) to handle WhatsApp conversations, qualify leads, trigger automations and move deals through configurable pipelines. Multi-tenant with LGPD-minded controls and a one-command VPS installer for full data ownership.

NuExtract3

Unified 4B vision-language model for document understanding that converts images or text into template-driven structured JSON or clean Markdown. Key features: multimodal inputs (image+text), template-based extraction, reasoning vs non-reasoning modes, and vLLM/OpenAI-compatible deployment for OCR, invoice/forms extraction, and RAG preprocessing.

Knowledge Catalog

Provides tools and samples to build context management, enrichment, and retrieval solutions on Google Cloud Knowledge Catalog — an AI-oriented data catalog that builds a dynamic knowledge graph for structured and unstructured data, suitable for RAG and agent workflows.

TiniX Vietnam OCR Annual Financial Statements

OCR-extracted Vietnamese annual financial reports (2015–2025) from 18,231 filings across 1,491 tickers — plain-text OCR outputs for document-QA, information extraction, VLM/RAG development. Contains only TXT OCR files; CC BY-NC 4.0 license.

Masking Stale Observations Helps Search Agents -- Until It Doesn't: A Regime Map and Its Mechanism

Analyzes when masking stale observations improves long-horizon search agents and why, identifying an asymmetric inverted-U relationship between masking benefit, retriever quality, and model capacity; explains a token-for-turn trade-off and releases evaluation scaffolds and trajectories.

Memory is Reconstructed, Not Retrieved: Graph Memory for LLM Agents

Represents episodic memory as a Cue–Tag–Content graph and integrates LLM reasoning into active retrieval so agents iteratively reconstruct and prune evidence paths for long-horizon questions. Reports up to 23% gains on LoCoMo / LongMemEval while reducing token and runtime costs.

Code2LoRA: Hypernetwork-Generated Adapters for Code Language Models under Software Evolution

Generates repository-specific LoRA adapters via a hypernetwork to inject repo-level knowledge into code LMs with zero inference-time token overhead. Provides a Static snapshot mode and an Evo mode that updates adapters per commit; evaluated on the 604-repo RepoPeftBench.

NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16

Open-weight frontier LLM for agentic reasoning and long-context analysis (up to 1M tokens). Uses a LatentMoE + Mamba-2 hybrid with Multi-Token Prediction and NVFP4 efficiency (550B total / 55B active). Suited for multilingual agents, RAG, and heavy tool-use workloads.

MemDreamer: Decoupling Perception and Reasoning for Long Video Understanding via Hierarchical Graph Memory and Agentic Retrieval Mechanism

Decouples perception and reasoning for hours-long videos by streaming inputs into a three-tier Hierarchical Graph Memory and using an agentic Observation–Reason–Action retrieval loop; reduces reasoning context to ~2% of full video while improving benchmark accuracy.

LEDGER — Long-Context KPI Question Answering & Page Retrieval

Provides page-level relevance judgments and full OCR'd annual-report text for KPI question answering and page retrieval benchmarking — supports retrieval (per-page qrels) and needle‑in‑a‑haystack numeric extraction over long documents, with eval and train configs.

Rethinking RAG in Long Videos: What to Retrieve and How to Use It?

Proposes chunk-level multimodal retrieval and chunk-adaptive reranking for retrieval-augmented generation on long egocentric videos; introduces V-RAGBench to decouple retrieval vs. generation evaluation and CARVE to run parallel retrievers and select per-chunk configurations.