Pith. sign in

REVIEW 31 cited by

LLM2Vec: Large Language Models Are Secretly Powerful Text Encoders

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 2404.05961 v2 pith:NZXYC7PV submitted 2024-04-09 cs.CL cs.AI

LLM2Vec: Large Language Models Are Secretly Powerful Text Encoders

classification cs.CL cs.AI
keywords modelsllm2vectexttaskslargellmsstate-of-the-artunsupervised
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
read the original abstract

Large decoder-only language models (LLMs) are the state-of-the-art models on most of today's NLP tasks and benchmarks. Yet, the community is only slowly adopting these models for text embedding tasks, which require rich contextualized representations. In this work, we introduce LLM2Vec, a simple unsupervised approach that can transform any decoder-only LLM into a strong text encoder. LLM2Vec consists of three simple steps: 1) enabling bidirectional attention, 2) masked next token prediction, and 3) unsupervised contrastive learning. We demonstrate the effectiveness of LLM2Vec by applying it to 4 popular LLMs ranging from 1.3B to 8B parameters and evaluate the transformed models on English word- and sequence-level tasks. We outperform encoder-only models by a large margin on word-level tasks and reach a new unsupervised state-of-the-art performance on the Massive Text Embeddings Benchmark (MTEB). Moreover, when combining LLM2Vec with supervised contrastive learning, we achieve state-of-the-art performance on MTEB among models that train only on publicly available data (as of May 24, 2024). Our strong empirical results and extensive analysis demonstrate that LLMs can be effectively transformed into universal text encoders in a parameter-efficient manner without the need for expensive adaptation or synthetic GPT-4 generated data.

discussion (0)

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

Forward citations

Cited by 31 Pith papers

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

  1. 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.

  2. Lost in a Single Vector: Improving Long-Document Retrieval with Chunk Evidence Aggregation

    cs.CL 2026-06 unverdicted novelty 7.0

    DICE aggregates independently encoded document chunks into a single vector to reduce evidence dilution in long-document dense retrieval, reporting gains on LongEmbed especially beyond 4k tokens.

  3. OmniRetriever: Any-to-Any Audio-Video-Text Retrieval via Fusion-as-Teacher Distillation

    cs.CV 2026-05 unverdicted novelty 7.0

    OmniRetriever-7B uses fusion-as-teacher distillation plus Tuple-InfoNCE to improve any-to-any audio-video-text retrieval over prior open and closed models.

  4. Think When Needed: Adaptive Reasoning-Driven Multimodal Embeddings with a Dual-LoRA Architecture

    cs.CV 2026-05 unverdicted novelty 7.0

    TWN attaches separate reasoning and embedding LoRA adapters to a frozen backbone with gradient detachment and a self-supervised gate that decides per input whether to generate CoT, achieving SOTA on MMEB-V2 with 3-5% ...

  5. Latent Abstraction for Retrieval-Augmented Generation

    cs.CL 2026-04 unverdicted novelty 7.0

    LAnR unifies retrieval-augmented generation inside a single LLM by deriving dense retrieval vectors from a [PRED] token's hidden states and using entropy to adaptively stop retrieval, outperforming prior RAG on six QA...

  6. mEOL: Training-Free Instruction-Guided Multimodal Embedder for Vector Graphics and Image Retrieval

    cs.CV 2026-04 unverdicted novelty 7.0

    mEOL creates aligned embeddings for text, images, and SVGs using instruction-guided MLLM one-word summaries and semantic SVG rewriting, outperforming baselines on a new text-to-SVG retrieval benchmark.

  7. Can LLMs Deobfuscate Binary Code? A Systematic Analysis of Large Language Models into Pseudocode Deobfuscation

    cs.SE 2026-04 unverdicted novelty 7.0

    LLM deobfuscation of binaries to pseudocode depends more on reasoning ability and task-specific fine-tuning than on model size, with reasoning models showing robustness across ISAs and obfuscation levels on the new Bi...

  8. VLM2Vec: Training Vision-Language Models for Massive Multimodal Embedding Tasks

    cs.CV 2024-10 conditional novelty 7.0

    VLM2Vec converts state-of-the-art vision-language models into universal multimodal embedders via contrastive training on the new MMEB benchmark, delivering 10-20% absolute gains over prior models on both in-distributi...

  9. IRIS: Reusable Identity Representations from Frozen LLMs for Entity Alignment

    cs.CL 2026-07 conditional novelty 6.0

    IRIS extracts identity embeddings from frozen LLMs so each entity is encoded once from its own knowledge graph and matched to other graphs by cosine similarity, hitting 97.99-100.00 Hits@1 on four benchmarks.

  10. 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.

  11. GraspLLM: Towards Zero-Shot Generalization on Text-Attributed Graphs with LLMs

    cs.CL 2026-06 unverdicted novelty 6.0

    GraspLLM extracts dataset-agnostic structural patterns via motif contrastive learning and aligns contextual subgraphs to LLM tokens, outperforming prior LLM-based methods on TAGs especially in zero-shot settings.

  12. Your UnEmbedding Matrix is Secretly a Feature Lens for Text Embeddings

    cs.CL 2026-06 unverdicted novelty 6.0

    EmbedFilter applies a linear filter derived from the LLM unembedding matrix to suppress high-frequency token influences in text embeddings, yielding improved zero-shot performance and inherent dimensionality reduction.

  13. Fine-grained Fragment Retrieval in Multi-modal Long-form Dialogues

    cs.CL 2026-06 unverdicted novelty 6.0

    Introduces FFR task, F2RVLM and FFRS models, and MLDR dataset for retrieving coherent multi-modal dialogue fragments, reporting superior performance on single-dialogue and corpus benchmarks.

  14. On the Robustness of Multilingual Text Embedding Rankings Across Learning Tasks, Languages, and Benchmark Datasets

    cs.CL 2026-05 unverdicted novelty 6.0

    Meta-study of MTEB rankings introduces dataset-composition and ranking-scheme robustness indicators and finds only a small subset of models stay consistently strong across tasks, languages, and evaluation variations.

  15. Anticipating Innovation Using Large Language Models

    cs.CL 2026-05 unverdicted novelty 6.0

    TechToken uses transformer embeddings of IPC codes to measure linguistic convergence in patents and predict future technological combinations.

  16. Interpretable Difficulty-Aware Knowledge Tracing in Tutor-Student Dialogues

    cs.CL 2026-05 unverdicted novelty 6.0

    A difficulty-aware conversational knowledge tracing framework that combines LLMs with Item Response Theory to produce interpretable student performance predictions in tutor dialogues.

  17. Modular Representation Compression: Adapting LLMs for Efficient and Effective Recommendations

    cs.IR 2026-04 unverdicted novelty 6.0

    LLMs exhibit mid-layer representation advantage for recommendations; MARC compresses representations modularly to reduce costs while improving performance, as shown in a large-scale online advertising deployment.

  18. RePrompT: Recurrent Prompt Tuning for Integrating Structured EHR Encoders with Large Language Models

    cs.CL 2026-04 unverdicted novelty 6.0

    RePrompT uses recurrent prompt tuning to inject prior-visit latent states and cohort-derived population prompt tokens into LLMs, yielding better performance than pure EHR or pure LLM baselines on MIMIC clinical predic...

  19. Turning Generators into Retrievers: Unlocking MLLMs for Natural Language-Guided Geo-Localization

    cs.CV 2026-04 unverdicted novelty 6.0

    Parameter-efficient fine-tuning lets MLLMs serve as effective retrievers for natural-language-guided cross-view geo-localization, beating dual-encoder baselines on GeoText-1652 and CVG-Text while using far fewer train...

  20. InsTraj: Instructing Diffusion Models with Travel Intentions to Generate Real-world Trajectories

    cs.AI 2026-04 unverdicted novelty 6.0

    InsTraj generates realistic, instruction-faithful GPS trajectories by using an LLM to parse natural-language travel intent and a multimodal diffusion transformer to produce the paths.

  21. Conjuring Semantic Similarity

    cs.AI 2024-10 unverdicted novelty 6.0

    Semantic similarity between texts is measured by the Jeffreys divergence between the image distributions induced by conditioning a diffusion model on each text, computed via Monte-Carlo sampling of the reverse-time SDEs.

  22. 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.

  23. Probe, Don't Prompt: A Hidden-State Probe for Metadata Filtering in Multi-Meta-RAG

    cs.CL 2026-07 conditional novelty 5.0

    A shallow mean-pooled multi-label probe on a 135M–1.5B open model replaces GPT-3.5 source extraction in Multi-Meta-RAG at 90.9% set-exact accuracy, with the margin coming from null-query abstention.

  24. MedMIX: Modality-Internal Expert Fusion for Multimodal Medical Diagnosis

    cs.LG 2026-05 unverdicted novelty 5.0

    MedMIX combines intra-modality expert fusion, learned inter-modality fusion, and training-only large-small collaboration to deliver robust multimodal medical prediction under incomplete modalities across three benchmarks.

  25. AFMRL: Attribute-Enhanced Fine-Grained Multi-Modal Representation Learning in E-commerce

    cs.CL 2026-04 unverdicted novelty 5.0

    AFMRL uses MLLM-generated attributes in attribute-guided contrastive learning and retrieval-aware reinforcement to achieve SOTA fine-grained multimodal retrieval on e-commerce datasets.

  26. 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.

  27. Smarter, Better, Faster, Longer: A Modern Bidirectional Encoder for Fast, Memory Efficient, and Long Context Finetuning and Inference

    cs.CL 2024-12 unverdicted novelty 5.0

    ModernBERT is a new bidirectional encoder model achieving SOTA performance on diverse classification and retrieval benchmarks while offering superior speed and memory efficiency for long-context inference.

  28. HARNESS-LM: A Three-Phase Training Recipe for Harnessing SLMs in Sponsored Search Retrieval

    cs.IR 2026-05 unverdicted novelty 4.0

    HARNESS-LM uses teacher fine-tuning, L2 query alignment, and contrastive refinement to distill large SLM retrievers into compact models that recover 98% precision with up to 27x lower latency on Bing Ads benchmarks.

  29. Towards Generalizable and Efficient Large-Scale Generative Recommenders

    cs.IR 2026-05 unverdicted novelty 4.0

    Empirical scaling of generative recommenders to 1B parameters shows task-dependent gains and requires targeted adaptations for production constraints like latency and item freshness.

  30. From Ambiguity to Accuracy: The Transformative Effect of Coreference Resolution on Retrieval-Augmented Generation systems

    cs.CL 2025-07 unverdicted novelty 4.0

    Coreference resolution improves retrieval relevance and QA performance in RAG systems, with mean pooling performing best and smaller models benefiting more.

  31. 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.