Knowledge Packs deliver knowledge via pre-computed KV caches with exact equivalence under causal masking, achieving zero divergences on tested questions and enabling value-based steering without training.
hub Canonical reference
C-Pack: Packed Resources For General Chinese Embeddings
Canonical reference. 75% of citing Pith papers cite this work as background.
80
Pith papers citing it
71
external citations · Pith
Background
75% of classified citations
abstract
We introduce C-Pack, a package of resources that significantly advance the field of general Chinese embeddings. C-Pack includes three critical resources. 1) C-MTEB is a comprehensive benchmark for Chinese text embeddings covering 6 tasks and 35 datasets. 2) C-MTP is a massive text embedding dataset curated from labeled and unlabeled Chinese corpora for training embedding models. 3) C-TEM is a family of embedding models covering multiple sizes. Our models outperform all prior Chinese text embeddings on C-MTEB by up to +10% upon the time of the release. We also integrate and optimize the entire suite of training methods for C-TEM. Along with our resources on general Chinese embedding, we release our data and models for English text embeddings. The English models achieve state-of-the-art performance on MTEB benchmark; meanwhile, our released English data is 2 times larger than the Chinese data. All these resources are made publicly available at https://github.com/FlagOpen/FlagEmbedding.
hub tools
citation-role summary
background 6
baseline 1
method 1
citation-polarity summary
representative citing papers
STEB is a new benchmark of 96 datasets in 7 languages for evaluating style text embeddings on authorship, detection, and linguistic probing tasks.
Cortex uses an Ontological Corpus Graph to structure web-scale corpora, creating a refined 24.14B-token corpus and a new benchmark validated on eight LLMs.
Retrieval coverage limits LLM rerankers in cold-start recommendation; a learned hybrid fusion improves pool quality but LLM reranking often degrades end-to-end performance while simpler rankers exploit the pool.
SkillWeaver formalizes compositional skill routing for LLM agents and introduces SAD, which raises step-level decomposition accuracy from 51% to 67.7% on a new 300-query benchmark over 2209 real MCP skills.
MetaSyn is a stage-level benchmark of 442 meta-analyses showing LLM agents retrieve up to 90.9% of eligible studies but include at most 52.7% in their final reports.
RRDA introduces a router plus separate edit and locality adapters for memory-assisted knowledge editing, reporting highest accuracy on CounterFact, ZsRE, and MQuAKE-CF across two 8B models.
LEDGER provides a corpus of 4,999 annual reports with 31 labeled KPIs and three benchmarks for page-level retrieval, needle-in-haystack lookup, and full KPI extraction from long documents.
Proves Voronoi complexity equals sign-rank for top-1 retrieval, introduces CUS diagnostic predicting retrieval failure at AUC >0.8 without labels, and AT-DW-InfoNCE objective with derived alpha^*=2.0 that improves Recall@100 on synthetic data.
ICICLE is an in-context indexing method for generative retrieval that uses source-aware docid generation with [COPY] routing and calibration to handle new documents without retraining.
IdioLink introduces a benchmark dataset and evaluation showing that strong embedding models struggle to retrieve equivalent meanings across idiomatic and literal forms, relying on shallow cues instead.
Prism-Reranker models output relevance, contribution statements, and evidence passages to support agentic retrieval beyond scalar scoring.
HaS accelerates RAG retrieval via homology-aware speculative retrieval and homologous query re-identification validation, cutting latency 24-37% with 1-2% accuracy drop on tested datasets.
METRO induces both short-term actions and long-term planning from expert transcripts into a Strategy Forest, outperforming prior methods by 9-10% on two non-collaborative dialogue benchmarks.
PERMA is a new benchmark using temporally ordered events, text variability, and linguistic alignment to evaluate LLM memory agents on persona consistency beyond simple retrieval.
CHIMERA is the first large-scale mined KB of concept recombinations from scientific literature, created via a new IE task and LLM extraction, with demonstrated uses in pattern analysis and hypothesis generation.
VisRAG achieves 20-40% better end-to-end performance than text-based RAG by directly embedding and retrieving document images with VLMs.
MultiHop-RAG is a new benchmark dataset demonstrating that existing retrieval-augmented generation systems perform poorly on multi-hop queries requiring retrieval and reasoning over multiple evidence pieces.
Structured retrieval (NAVEMBED/NAVINDEX) matches full-corpus injection on legal multi-document QA while cutting tokens 1.6–30×, with a closed-form cache-crossover cost rule.
Graph-PRefLexOR fine-tunes graph-native models with GRPO to organize reasoning into phases, yielding 40-65% gains in traceable hypothesis generation and 2-3x semantic diversity on 100 materials science questions.
Permutation-invariant fine-tuning (PI-FT) randomizes field order and applies dropout during embedding model training to eliminate sensitivity to serialization order, reducing order-change penalty from 7.4 to 0.2 nDCG@10 on a generated multilingual DevDataBench while outperforming zero-shot baselines
GeoRAG recasts RAG context selection as monotone submodular Information Demand Coverage Optimization solved via Sinkhorn-Wasserstein distance, delivering +6.5 to +7.5 EM gains over top-k on six QA benchmarks.
A memory that keeps a wrong conclusion while dropping the source makes an LLM confidently repeat the error; keeping the source instead lets a later correction recover the truth at the same budget.
Presents Invoice Haystack benchmark for homogeneous document retrieval and VL-RAG hybrid framework achieving 60% Recall@1 and up to 13.5 point gains over prior methods.
citing papers explorer
-
Knowledge Packs: Zero-Token Knowledge Delivery via KV Cache Injection
Knowledge Packs deliver knowledge via pre-computed KV caches with exact equivalence under causal masking, achieving zero divergences on tested questions and enabling value-based steering without training.
-
STEB: Style Text Embedding Benchmark
STEB is a new benchmark of 96 datasets in 7 languages for evaluating style text embeddings on authorship, detection, and linguistic probing tasks.
-
CORTEX: High-Quality Cross-Domain Organization of Web-Scale Corpora through Ontological Corpus Graph
Cortex uses an Ontological Corpus Graph to structure web-scale corpora, creating a refined 24.14B-token corpus and a new benchmark validated on eight LLMs.
-
Diagnosing and Mitigating Retrieval Bottlenecks in LLM-Based Cold-Start Recommendation
Retrieval coverage limits LLM rerankers in cold-start recommendation; a learned hybrid fusion improves pool quality but LLM reranking often degrades end-to-end performance while simpler rankers exploit the pool.
-
Compositional Skill Routing for LLM Agents: Decompose, Retrieve, and Compose
SkillWeaver formalizes compositional skill routing for LLM agents and introduces SAD, which raises step-level decomposition accuracy from 51% to 67.7% on a new 300-query benchmark over 2209 real MCP skills.
-
MetaSyn: A Benchmark for LLM Agents on Meta-Analysis Articles from Nature Portfolio
MetaSyn is a stage-level benchmark of 442 meta-analyses showing LLM agents retrieve up to 90.9% of eligible studies but include at most 52.7% in their final reports.
-
When to Write and When to Suppress: Route-Specialized Dual Adapters for Memory-Assisted Knowledge Editing
RRDA introduces a router plus separate edit and locality adapters for memory-assisted knowledge editing, reporting highest accuracy on CounterFact, ZsRE, and MQuAKE-CF across two 8B models.
-
LEDGER: A Long-Context Benchmark of Corporate Annual Reports for Grounded Financial Retrieval and Extraction
LEDGER provides a corpus of 4,999 annual reports with 31 labeled KPIs and three benchmarks for page-level retrieval, needle-in-haystack lookup, and full KPI extraction from long documents.
-
The Voronoi Bottleneck: Capacity-Aware Dense Retrieval for Product Search
Proves Voronoi complexity equals sign-rank for top-1 retrieval, introduces CUS diagnostic predicting retrieval failure at AUC >0.8 without labels, and AT-DW-InfoNCE objective with derived alpha^*=2.0 that improves Recall@100 on synthetic data.
-
ICICLE: Expanding Retrieval with In-Context Documents
ICICLE is an in-context indexing method for generative retrieval that uses source-aware docid generation with [COPY] routing and calibration to handle new documents without retraining.
-
IdioLink: Retrieving Meaning Beyond Words Across Idiomatic and Literal Expressions
IdioLink introduces a benchmark dataset and evaluation showing that strong embedding models struggle to retrieve equivalent meanings across idiomatic and literal forms, relying on shallow cues instead.
-
Prism-Reranker: Beyond Relevance Scoring -- Jointly Producing Contributions and Evidence for Agentic Retrieval
Prism-Reranker models output relevance, contribution statements, and evidence passages to support agentic retrieval beyond scalar scoring.
-
HaS: Accelerating RAG through Homology-Aware Speculative Retrieval
HaS accelerates RAG retrieval via homology-aware speculative retrieval and homologous query re-identification validation, cutting latency 24-37% with 1-2% accuracy drop on tested datasets.
-
METRO: Towards Strategy Induction from Expert Dialogue Transcripts for Non-collaborative Dialogues
METRO induces both short-term actions and long-term planning from expert transcripts into a Strategy Forest, outperforming prior methods by 9-10% on two non-collaborative dialogue benchmarks.
-
PERMA: Benchmarking Personalized Memory Agents via Event-Driven Preference and Realistic Task Environments
PERMA is a new benchmark using temporally ordered events, text variability, and linguistic alignment to evaluate LLM memory agents on persona consistency beyond simple retrieval.
-
CHIMERA: A Knowledge Base of Scientific Idea Recombinations for Research Analysis and Ideation
CHIMERA is the first large-scale mined KB of concept recombinations from scientific literature, created via a new IE task and LLM extraction, with demonstrated uses in pattern analysis and hypothesis generation.
-
VisRAG: Vision-based Retrieval-augmented Generation on Multi-modality Documents
VisRAG achieves 20-40% better end-to-end performance than text-based RAG by directly embedding and retrieving document images with VLMs.
-
MultiHop-RAG: Benchmarking Retrieval-Augmented Generation for Multi-Hop Queries
MultiHop-RAG is a new benchmark dataset demonstrating that existing retrieval-augmented generation systems perform poorly on multi-hop queries requiring retrieval and reasoning over multiple evidence pieces.
-
Inject or Navigate? Token-Efficient Retrieval for LLM Analysis of Transactional Legal Documents
Structured retrieval (NAVEMBED/NAVINDEX) matches full-corpus injection on legal multi-document QA while cutting tokens 1.6–30×, with a closed-form cache-crossover cost rule.
-
Graph-Native Reinforcement Learning Enables Traceable Scientific Hypothesis Generation through Conceptual Recombination
Graph-PRefLexOR fine-tunes graph-native models with GRPO to organize reasoning into phases, yielding 40-65% gains in traceable hypothesis generation and 2-3x semantic diversity on 100 materials science questions.
-
Field Order Should Not Matter: Permutation-Invariant Embedding Model Fine-Tuning for Structured Metadata Retrieval
Permutation-invariant fine-tuning (PI-FT) randomizes field order and applies dropout during embedding model training to eliminate sensitivity to serialization order, reducing order-change penalty from 7.4 to 0.2 nDCG@10 on a generated multilingual DevDataBench while outperforming zero-shot baselines
-
Covering the Unseen: Information Demand Coverage Optimization for Retrieval-Augmented Generation
GeoRAG recasts RAG context selection as monotone submodular Information Demand Coverage Optimization solved via Sinkhorn-Wasserstein distance, delivering +6.5 to +7.5 EM gains over top-k on six QA benchmarks.
-
Reclaim Evaluation: A Lossy Memory Is Worse Than an Empty One
A memory that keeps a wrong conclusion while dropping the source makes an LLM confidently repeat the error; keeping the source instead lets a later correction recover the truth at the same budget.
-
Invoice Haystack: Benchmarking Document Retrieval and Visual Question Answering Under Strong Visual Homogeneity
Presents Invoice Haystack benchmark for homogeneous document retrieval and VL-RAG hybrid framework achieving 60% Recall@1 and up to 13.5 point gains over prior methods.
-
Scaling Dense Retrieval with LLM-Annotated Training Data: Structured Mining and Progressive Curriculum for E-Commerce Sponsored Search
A pipeline using multi-system retrieval disagreement, LLM-graded annotation at 89.1% human agreement, and staged curriculum on 240M+ examples trains a two-tower model that improves NDCG@10 by 5.1% and online ad metrics in Walmart sponsored search.
-
When Global Gating Is Enough: Admission-Time Hubness Control in Anisotropic Vector Retrieval Systems
Global admission-time gating with sentinel queries controls vector hubness in anisotropic embeddings, achieving recall 1.0 on critical attack points and 0.91 on HotFlip attacks with 1% false positives, while per-topic gates add no benefit.
-
Less Context, More Accuracy: A Bi-Temporal Memory Engine for LLM Agents Where a Lean Retrieved Context Beats the Full History
Engram's hybrid bi-temporal retrieval from a knowledge graph with provenance yields 83.6% accuracy on LongMemEval_S using 9.6k tokens versus 73.2% with full history.
-
RWGBench: Evaluating Scholarly Positioning in Related Work Generation
RWGBench measures related-work generation by citation choices, and shows citation-focused metrics expose failures that text-similarity and LLM-judge scores miss.
-
Anchorless Diversification for Parallel LLM Ideation
Semantic direction stratification yields the best diversity-quality-compute frontier for parallel LLM ideation across creative tasks.
-
The Routing Plateau: Understanding and Breaking the Accuracy Limits of LLM Routers
LLM routers across 21 methods on 5 benchmarks converge to similar accuracy below oracle due to learning global performance trends rather than fine-grained query signals.
-
Structure Retention in Embedding Spaces as a Predictor of Benchmark Performance
Embedding model performance on MTEB tasks correlates strongly with nearest-neighbor overlap and ICA magnitude differences in their embedding spaces.
-
An Annotation Scheme and Classifier for Personal Facts in Dialogue
An extended annotation scheme with new categories and attributes plus a Gemma-300M-based multi-head classifier achieves 81.6% macro F1 on personal fact classification, outperforming few-shot LLM baselines by nearly 9 points with lower compute.
-
SkillRAE: Agent Skill-Based Context Compilation for Retrieval-Augmented Execution
SkillRAE organizes skills into a graph and compiles compact, grounded contexts for LLM agents, yielding 11.7% gains on SkillsBench over prior RAE methods.
-
Retrieval from Within: An Intrinsic Capability of Attention-Based Models
Attention-based models can retrieve evidence intrinsically by using decoder attention to score and reuse their own pre-encoded chunks, outperforming separate retrieval pipelines on QA benchmarks.
-
A Replicability Study of XTR
XTR training does not improve retrieval effectiveness over ColBERT but enhances IVF engine efficiency by flattening token scores to produce more discriminative centroids.
-
MemRouter: Memory-as-Embedding Routing for Long-Term Conversational Agents
A lightweight supervised router using frozen-LLM embeddings for memory admission decisions outperforms LLM-based memory managers in both F1 score and latency on the LoCoMo benchmark.
-
MiMIC: Mitigating Visual Modality Collapse in Universal Multimodal Retrieval While Avoiding Semantic Misalignment
MiMIC mitigates visual modality collapse and semantic misalignment in universal multimodal retrieval via fusion-in-decoder architecture and robust single-modality training.
-
EvoRAG: Making Knowledge Graph-based RAG Automatically Evolve through Feedback-driven Backpropagation
EvoRAG adds a feedback-driven backpropagation step that attributes response quality to individual knowledge-graph triplets and updates the graph to raise reasoning accuracy by 7.34 percent over prior KG-RAG methods.
-
Regime-Conditional Retrieval: Theory and a Transferable Router for Two-Hop QA
Two-hop QA retrieval performance depends on whether the hop-2 entity is in the question or bridge passage, and a simple predicate-based router trained on one dataset transfers to improve R@5 on others.
-
ResearchEVO: An End-to-End Framework for Automated Scientific Discovery and Documentation
ResearchEVO automates the discover-then-explain cycle by evolving algorithms via fitness-driven LLM co-evolution and generating grounded, anti-hallucination research papers through sentence-level RAG.
-
SelRoute: Query-Type-Aware Routing for Long-Term Conversational Memory Retrieval
SelRoute routes queries to type-specific retrieval pipelines, achieving Recall@5 of 0.800 with a 109M model on LongMemEval_M and outperforming LLM-augmented baselines including a strong zero-ML lexical method.
-
ASTRA: Mapping Art-Technology Institutions via Conceptual Axes, Text Embeddings, and Unsupervised Clustering
ASTRA combines an eight-axis conceptual framework with text embeddings and unsupervised clustering to map and group 78 art-technology institutions into coherent thematic clusters.
-
LLM-AutoDP: Automatic Data Processing via LLM Agents for Model Fine-tuning
LLM agents iteratively generate and optimize data processing strategies for fine-tuning, delivering over 80% win rates versus unprocessed data and 65% versus LLM-based AutoML baselines while cutting search time by up to 10x.
-
Retrieval-Augmented Generation for Natural Language Processing: A Survey
The survey organizes RAG methods via a taxonomy of query-based, logits-based, latent, and parametric fusion with comparisons on accessibility, efficiency, applications, and challenges.
-
Crystallizing Schemas with Teleoscope: Thematic Curation of Large Text Corpora on Reddit
Teleoscope enables thematic curation of large Reddit corpora via interactive refinement, with three deployments indicating benefits in serendipitous keyword discovery, search saturation confidence, and collaborative curation discussions.
-
Multimodal Video-to-Music Recommendation via Semantic Retrieval and Temporal Reranking
VTMR is a two-stage video-to-music recommender: joint audio-visual-text retrieval of candidates, then temporal-sequence reranking, lifting R@10 to 18.3 and matching commercial preference.
-
Dissecting Agentic RAG: A Component Ablation for Multi-Hop QA with a Local 7B Model
Ablation on local 7B agentic RAG shows fixed hybrid retrieval outperforms adaptive routing and two iterations capture nearly all gains from five on HotpotQA.
-
Stellar: Scalable Multimodal Document Retrieval for Natural Language Queries
Stellar reduces memory and latency by 1-2 orders of magnitude for multimodal document retrieval via lexical filtering with a fine-tuned MLLM and disk-backed late interaction without loss in effectiveness.
-
Auditing Proprietary Alignment in Large Language Models: A Comparative Framework Without a Ground-Truth Standard
A comparative statistical framework is proposed to audit proprietary alignment in black-box LLMs by quantifying behavioral divergences from reference models rather than absolute correctness.
-
DMF: A Deterministic Memory Framework for Conversational AI Agents
DMF introduces a deterministic memory system using Survival Scores and decay laws that matches Mem0 accuracy on benchmarks while eliminating LLM token use for memory preparation.