Pith. sign in

REVIEW 4 cited by

LongEmbed: Extending Embedding Models for Long Context Retrieval

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 2404.12096 v3 pith:O6ABGUAR submitted 2024-04-18 cs.CL cs.LG

LongEmbed: Extending Embedding Models for Long Context Retrieval

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

Embedding models play a pivot role in modern NLP applications such as IR and RAG. While the context limit of LLMs has been pushed beyond 1 million tokens, embedding models are still confined to a narrow context window not exceeding 8k tokens, refrained from application scenarios requiring long inputs such as legal contracts. This paper explores context window extension of existing embedding models, pushing the limit to 32k without requiring additional training. First, we examine the performance of current embedding models for long context retrieval on our newly constructed LongEmbed benchmark. LongEmbed comprises two synthetic tasks and four carefully chosen real-world tasks, featuring documents of varying length and dispersed target information. Benchmarking results underscore huge room for improvement in these models. Based on this, comprehensive experiments show that training-free context window extension strategies like position interpolation can effectively extend the context window of existing embedding models by several folds, regardless of their original context being 512 or beyond 4k. Furthermore, for models employing absolute position encoding (APE), we show the possibility of further fine-tuning to harvest notable performance gains while strictly preserving original behavior for short inputs. For models using rotary position embedding (RoPE), significant enhancements are observed when employing RoPE-specific methods, such as NTK and SelfExtend, indicating RoPE's superiority over APE for context window extension. To facilitate future research, we release E5-Base-4k and E5-RoPE-Base, along with the LongEmbed benchmark.

discussion (0)

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

Forward citations

Cited by 4 Pith papers

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

  1. Bekko Embedding: Parameter-Efficient Multilingual Retrieval with Ultra-Compact Encoders

    cs.IR 2026-07 conditional novelty 7.0

    Bekko a8m, with 7.7M active parameters, scores 56.2 on MMTEB Multilingual v2 Retrieval, beating mE5 models and BGE-M3, while a25m reaches 57.5, on par with gte-multilingual-base.

  2. HAKARI-Bench: A Lightweight Benchmark for Comparing Retrieval Architectures and Efficiency Settings under Unified Conditions

    cs.IR 2026-06 unverdicted novelty 7.0

    HAKARI-Bench reconstructs 35 benchmarks into 551 tasks across 43 languages, reproducing full MTEB, MMTEB, and BEIR rankings with Spearman correlation above 0.97 while supporting efficiency variant comparisons.

  3. EmbeddingGemma: Powerful and Lightweight Text Representations

    cs.CL 2025-09 unverdicted novelty 6.0

    A 300M-parameter open embedding model sets new SOTA on MTEB for its size class and matches models twice as large while staying effective when compressed.

  4. 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.