Pith. sign in

REVIEW 39 cited by

Improving Text Embeddings with Large Language Models

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2401.00368 v3 pith:YOMD7RN5 submitted 2023-12-31 cs.CL cs.IR

Improving Text Embeddings with Large Language Models

classification cs.CL cs.IR
keywords datatextsyntheticlabeledmethodbenchmarksdatasetsembedding
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
Share X Bluesky LinkedIn Reddit HN
read the original abstract

In this paper, we introduce a novel and simple method for obtaining high-quality text embeddings using only synthetic data and less than 1k training steps. Unlike existing methods that often depend on multi-stage intermediate pre-training with billions of weakly-supervised text pairs, followed by fine-tuning with a few labeled datasets, our method does not require building complex training pipelines or relying on manually collected datasets that are often constrained by task diversity and language coverage. We leverage proprietary LLMs to generate diverse synthetic data for hundreds of thousands of text embedding tasks across 93 languages. We then fine-tune open-source decoder-only LLMs on the synthetic data using standard contrastive loss. Experiments demonstrate that our method achieves strong performance on highly competitive text embedding benchmarks without using any labeled data. Furthermore, when fine-tuned with a mixture of synthetic and labeled data, our model sets new state-of-the-art results on the BEIR and MTEB benchmarks.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 39 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. MathAtlas: A Benchmark for Autoformalization in the Wild

    cs.AI 2026-05 accept novelty 8.0

    MathAtlas is the first large-scale benchmark for autoformalizing graduate mathematics, where even strong models reach only 9.8% correctness on theorem statements and drop to 2.6% on the hardest dependency-deep subset.

  2. One Generator, Any Process: LLM-Conditioning for the LHC

    hep-ph 2026-06 unverdicted novelty 7.0

    LLM embeddings condition generative networks for LHC events, yielding faster convergence, higher quality, and generalization to unseen processes.

  3. Fine-grained Claim-level RAG Benchmark for Law

    cs.CL 2026-05 unverdicted novelty 7.0

    ClaimRAG-LAW is a French-English legal RAG benchmark with claim-level granularity for experts and non-experts that reveals limitations in current retrieval and generation performance.

  4. Fine-grained Claim-level RAG Benchmark for Law

    cs.CL 2026-05 unverdicted novelty 7.0

    ClaimRAG-LAW provides a claim-level bilingual benchmark and evaluation framework for legal RAG systems that reveals limitations in current retrieval and generation performance.

  5. Fine-grained Claim-level RAG Benchmark for Law

    cs.CL 2026-05 unverdicted novelty 7.0

    ClaimRAG-LAW is a new English-French dataset with diverse question types and a claim-level evaluation framework that reveals limitations in current legal RAG systems for retrieval, generation, and analysis.

  6. Semantic Reranking at Inference Time for Hard Examples in Rhetorical Role Labeling

    cs.CL 2026-05 unverdicted novelty 7.0

    RISE is an inference-time semantic reranking framework that refines low-confidence predictions in rhetorical role labeling using contrastively learned label representations, delivering an average +9.15 macro-F1 gain o...

  7. Test-Time Compute for Frozen Embedding Models through Agentic Program Search

    cs.LG 2026-05 unverdicted novelty 7.0

    Agentic program search over frozen embedding APIs yields a parameter-free inference algebra—a softmax-weighted centroid of top-K documents interpolated with the query—that lifts nDCG@10 across seven model families on ...

  8. Test-Time Compute for Frozen Embedding Models through Agentic Program Search

    cs.LG 2026-05 unverdicted novelty 7.0

    A softmax-weighted centroid of the local top-K documents interpolated with the query improves nDCG@10 for frozen embedding models across seven families on held-out BEIR data.

  9. IntrAgent: An LLM Agent for Content-Grounded Information Retrieval through Literature Review

    cs.IR 2026-04 unverdicted novelty 7.0

    IntrAgent uses a two-stage pipeline of section ranking and iterative reading to perform content-grounded literature information retrieval, achieving 13.2% higher accuracy than RAG and agent baselines on the new IntraB...

  10. Unified Work Embeddings: Contrastive Learning of a Bidirectional Multi-task Ranker

    cs.CL 2025-11 unverdicted novelty 7.0

    UWE is a task-agnostic bi-encoder that uses many-to-many InfoNCE and token-level soft late interaction to achieve zero-shot ranking across unseen work-related target spaces while using far fewer parameters than Qwen3-...

  11. Replacing Judges with Juries: Evaluating LLM Generations with a Panel of Diverse Models

    cs.CL 2024-04 conditional novelty 7.0

    A panel of smaller diverse LLMs outperforms a single large model as an evaluator of generations, showing less intra-model bias and over 7x lower cost.

  12. M3-Embedding: Multi-Linguality, Multi-Functionality, Multi-Granularity Text Embeddings Through Self-Knowledge Distillation

    cs.CL 2024-02 unverdicted novelty 7.0

    M3-Embedding is a single model for multi-lingual, multi-functional, and multi-granular text embeddings trained via self-knowledge distillation that achieves new state-of-the-art results on multilingual, cross-lingual,...

  13. UEmbed: Unified Sparse and Dense Multimodal Embeddings

    cs.CV 2026-08 conditional novelty 6.0

    UEmbed uses 16 special tokens over a partitioned vocabulary to make a decoder-only multimodal model emit dense and sparse embeddings in one forward pass; the 9B model scores 71.8 dense / 71.0 sparse on MMEB-v2.

  14. Bifrost: Empowering Pretrained Language Model with Fallibility Representation for Log-Based Fault Diagnosis

    cs.SE 2026-07 conditional novelty 6.0

    Bifrost fine-tunes BART-Base on logs with three self-supervised contrastive tasks, beating general PLMs by roughly 10-20% on anomaly detection, root-cause localization, and fault identification.

  15. Universal Guideline-Driven Image Clustering via a Hybrid LLM Agent

    cs.CV 2026-06 unverdicted novelty 6.0

    A hybrid LLM agent framework performs universal image clustering by generating guideline-aware embeddings via concept proxies and using MST-based LLM traversal for automatic discovery.

  16. One Generator, Any Process: LLM-Conditioning for the LHC

    hep-ph 2026-06 unverdicted novelty 6.0

    LLM embeddings condition a generative transformer to enable faster convergence, better performance, and generalization to unseen LHC processes using a single model.

  17. Learning to Reason by Analogy via Retrieval-Augmented Reinforcement Fine-Tuning

    cs.CL 2026-06 unverdicted novelty 6.0

    RA-RFT trains a retriever to rank contexts by expected reasoning benefit and uses the retrieved analogies inside reinforcement fine-tuning, yielding 7.1 and 2.8 point gains on AIME 2025 over GRPO for two Qwen3 models.

  18. LaME: Learning to Think in Latent Space for Multimodal Embedding via Information Bottleneck

    cs.CV 2026-06 unverdicted novelty 6.0

    LaME performs latent multimodal embedding reasoning with K learnable reason tokens in a weakly supervised information bottleneck, matching some explicit CoT models while running 60x faster.

  19. LaME: Learning to Think in Latent Space for Multimodal Embedding via Information Bottleneck

    cs.CV 2026-06 conditional novelty 6.0

    Latent reasoning through eight bottleneck tokens gives multimodal retrieval performance close to explicit chain-of-thought models at a fraction of the inference cost.

  20. Semantic Retrieval for Product Search in E-Commerce

    cs.IR 2026-05 unverdicted novelty 6.0

    A dual-encoder LLM is trained via contrastive learning then ROAR to retrieve exact matches and rank substitutes in e-commerce search.

  21. Structure Retention in Embedding Spaces as a Predictor of Benchmark Performance

    cs.CL 2026-05 unverdicted novelty 6.0

    Embedding model performance on MTEB tasks correlates strongly with nearest-neighbor overlap and ICA magnitude differences in their embedding spaces.

  22. Test-Time Compute for Frozen Embedding Models through Agentic Program Search

    cs.LG 2026-05 unverdicted novelty 6.0

    Agentic program search over a frozen encoder API yields retrieval programs that improve nDCG@10 on held-out tasks and unseen encoder families with no per-domain training.

  23. Is Textual Similarity Invariant under Machine Translation? Evidence Based on the Political Manifesto Corpus

    cs.CL 2026-05 unverdicted novelty 6.0

    Machine translation preserves embedding similarity structure for ten languages but distorts it for four in the Manifesto Corpus, via a new non-inferiority testing framework.

  24. Regime-Conditional Retrieval: Theory and a Transferable Router for Two-Hop QA

    cs.IR 2026-04 conditional novelty 6.0

    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.

  25. E5-V: Universal Embeddings with Multimodal Large Language Models

    cs.CL 2024-07 unverdicted novelty 6.0

    E5-V produces strong universal multimodal embeddings from MLLMs trained solely on text pairs, often surpassing prior methods across retrieval and related tasks without multimodal fine-tuning.

  26. NV-Embed: Improved Techniques for Training LLMs as Generalist Embedding Models

    cs.CL 2024-05 accept novelty 6.0

    NV-Embed achieves first place on the MTEB leaderboard across 56 tasks by combining a latent attention layer, causal-mask removal, two-stage contrastive training, and data curation for LLM-based embedding models.

  27. Improving Long-Context Retrieval with Multi-Prefix Embedding

    cs.IR 2026-06 unverdicted novelty 5.0

    Multi-Prefix Embedding extracts per-chunk embeddings from a single forward pass over EOS-separated document chunks and matches via MaxSim while training only on document-level labels.

  28. Not All RAGs Are Created Equal: A Component-Wise Empirical Study for Software Engineering Tasks

    cs.SE 2026-05 unverdicted novelty 5.0

    Retriever-side choices, particularly the retrieval algorithm, exert more influence on RAG performance than generator selection across code generation, summarization, and repair tasks.

  29. Sustainable Hybrid Document-Routed Retrieval for Financial RAG: Resolving the Robustness-Precision Trade-off

    cs.CL 2026-03 unverdicted novelty 5.0

    HDRR combines document-level semantic routing with scoped chunk retrieval to outperform both pure chunk-based retrieval and semantic file routing on the FinDER benchmark, delivering higher average scores, lower failur...

  30. Sustainable Hybrid Document-Routed Retrieval for Financial RAG: Resolving the Robustness-Precision Trade-off

    cs.CL 2026-03 conditional novelty 5.0

    Hybrid Document-Routed Retrieval (HDRR) filters financial filings with LLM document routing then scopes chunk retrieval, beating pure chunk and pure file routing on FinDER accuracy and efficiency.

  31. Legal Retrieval for Public Defenders

    cs.IR 2026-01 conditional novelty 5.0

    NJ BriefBank is a domain-adapted legal retrieval tool for public defenders that improves on standard benchmarks by incorporating legal reasoning, domain data, and synthetic examples, with a new released taxonomy and a...

  32. Retrofitting Small Multilingual Models for Retrieval: Matching 7B Performance with 300M Parameters

    cs.CL 2025-10 conditional novelty 5.0

    A 300M multilingual embedding model matches or exceeds 7B retrieval performance via optimized data scale, hard negatives, and task diversity over language diversity.

  33. Improving Korean-English Cross-Lingual Retrieval: A Data-Centric Study of Language Composition and Model Merging

    cs.IR 2025-07 unverdicted novelty 5.0

    Language composition in training data creates opposing effects on CLIR and mono-IR performance for Korean-English retrieval, which model merging can partially resolve.

  34. VLM2Vec-V2: Advancing Multimodal Embedding for Videos, Images, and Visual Documents

    cs.CV 2025-07 unverdicted novelty 5.0

    VLM2Vec-V2 is a multimodal embedding model trained on an extended MMEB-V2 benchmark that adds video and visual document tasks and reports gains on both new and prior image benchmarks.

  35. Multilingual E5 Text Embeddings: A Technical Report

    cs.CL 2024-02 unverdicted novelty 5.0

    Open-source multilingual E5 embedding models are trained via contrastive pre-training on 1 billion text pairs and fine-tuning, with an instruction-tuned model matching English SOTA performance.

  36. Exploring Block Anomaly Detection In HDFS Log Data Analysis

    cs.LG 2026-07 conditional novelty 4.0

    An LLM-BiLSTM hybrid model detects anomalous HDFS log blocks with 94.8% F1 on the Loghub HDFS dataset, outperforming DeepLog.

  37. LPCAN: Lightweight Pyramid Cross-Attention Network for Rail Surface Defect Detection Using RGB-D Data

    cs.CV 2026-01 reject novelty 4.0

    A lightweight RGB-D cross-attention network is proposed for rail defect detection, but the SOTA accuracy and generalization claims are internally inconsistent and the implementation is not public.

  38. Are Decoder-Only Large Language Models the Silver Bullet for Code Search?

    cs.SE 2024-10 unverdicted novelty 4.0

    Fine-tuned decoder-only LLMs achieve up to 40.4% higher MAP than UniXcoder on CoSQA+ for code search, with non-monotonic size scaling and data composition sensitivity.

  39. Knowledge-Embedded and Hypernetwork-Guided Few-Shot Substation Meter Defect Image Generation Method

    cs.CV 2026-01 reject novelty 3.0

    Fine-tuning Stable Diffusion with DreamBooth-style knowledge and hypernetwork-guided crack control maps can synthesize substation meter defect images that boost a YOLOv8 defect detector's mAP when added to the training set.