Tag

ai-memory

Provides long-term memory for AI coding agents by compiling sanitized lifecycle observations into a git-versioned Markdown wiki that enables cross-agent handoffs, per-project isolation, and optional vector-backed retrieval.

Why Far Looks Up: Probing Spatial Representation in Vision-Language Models

Analyzes spatial representations in vision–language models and reveals a consistent vertical-position ↔ distance entanglement; introduces SpatialTunnel, a synthetic benchmark that exposes this perspective-driven shortcut, and provides code and a project page.

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.

Your UnEmbedding Matrix is Secretly a Feature Lens for Text Embeddings

Removes the subspace of frequent, uninformative tokens that LLMs inject into text embeddings via the model's unembedding matrix. EmbedFilter is a lightweight linear transform that refines LLM-derived embeddings to improve zero‑shot semantic retrieval, enable dimensionality reduction, and speed up indexing; code on GitHub.

TRL-Bench: Standardizing Cross-Paradigm Representation-Level Evaluation of Tabular Encoders

Standardizes representation-level evaluation for tabular encoders by exporting row-, column-, and table-level embeddings and probing them with shared lightweight heads across three suites (TRL-CTbench, TRL-Rbench, TRL-DLTE). Supplies curated benchmark assets and task rewrites (50 OpenML tables, 123 targets, a 47,772-table DLTE lake) to enable fair cross-paradigm comparison.

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.

AGE: Adaptive-masking for Graph Embedding in Graph Retrieval-Augmented Generation

Trains a transformer-based graph encoder with RL-guided adaptive masking so retrieved subgraphs embed relationships that better align with frozen LLM text encoders, improving GraphRAG performance with non-parametric retrievers on GraphQA benchmarks.

XERJ

Indexes mixed-format folders locally into a zero-config, Elasticsearch-compatible search engine offering keyword, semantic, vector and hybrid retrieval. Embeds on ingest, provides per-agent namespaced memory and AST-aware code indexing in a single static binary with no external embedding key.

Explorer_LLM_Rec_Competition

Provides anonymized multi-domain user behavior sequences and content metadata (short video, ads, e-commerce, live) for cross-domain recommendation, semantic-ID mapping, and content-understanding tasks. Key tables include per-user multi-domain behavior (~500k rows), pid→three-segment semantic IDs, captions, and level-3 tags; all item IDs are hashed for privacy.

Light-Omni: Reflex over Reasoning in Agentic Video Understanding with Long-Term Memory

Provides a reflexive agentic framework for long-horizon video understanding that replaces costly iterative reasoning with dual contextual states: a consolidated global multimodal script and parametric latent states for fast retrieval and response, improving speed and memory efficiency.

Octop

Self-hosted, multi-user, multi-agent AI assistant with a web dashboard, CLI, IM channel integrations, cron automation, RAG knowledge bases, and ACP IDE delegation. Runs as a single-process local-first service with pluggable workspace backends and plugin extensibility.

Dual Latent Memory in Vision-Language-Action Models for Robotic Manipulation

Reconstructs historical experience into latent memory tokens and weaves short- and long-term latent memories directly into vision-language-action reasoning to improve long-horizon robotic manipulation. Uses a four-part pipeline (curator, seeker, condenser, weaver) so memory participates natively in multimodal action formation.