KV-Embedding: Training-free Text Embedding via Internal KV Re-routing in Decoder-only LLMs (2026.acl-long)
Copied to clipboard
| Challenge: | Recent work shows that decoder-only LLMs can serve as strong embedding backbones when fine-tuned with contrastive objectives. |
| Approach: | They propose a framework that activates the latent representation power of frozen LLMs by rerouting the final token's KV states as a prepended prefix. |
| Outcome: | The proposed framework outperforms existing training-free baselines by 10% on MTEB and maintains robust performance on sequences up to 4,096 tokens. |
Similar Papers
A Systematic Study of Cross-Layer KV Sharing for Efficient LLM Inference (2025.naacl-short)
Copied to clipboard
| Challenge: | Recent studies have shown that sharing key-value (KV) cache across layers is effective in efficient inference of large language models. |
| Approach: | They propose a unified framework that covers several recent methods and their novel variants to investigate cross-layer KV sharing. |
| Outcome: | The proposed framework achieves higher throughput and better performance when reducing the size of the key-value cache by 2 while maintaining competitive performance. |
Causal2Vec: Improving Decoder-only LLMs as Embedding Models through a Contextual Token (2026.acl-long)
Copied to clipboard
| Challenge: | Existing methods modify attention mechanism to be bidirectional, undermining LLMs’ ability to extract semantic information acquired during pre-training. |
| Approach: | They propose a general-purpose embedding model that pre-encodes input text into a single Contextual token and then prepends it to the LLM's input sequence. |
| Outcome: | The proposed model improves performance of decoder-only large language models without altering their architectures or introducing significant computational overhead. |
GenEOL: Harnessing the Generative Power of LLMs for Training-Free Sentence Embeddings (2025.findings-naacl)
Copied to clipboard
| Challenge: | Training-free embedding methods focus on optimizing embeddable prompts . previous methods have overlooked the benefits of utilizing generative abilities of LLMs - GenEOL . |
| Approach: | They propose a method that leverages pretrained large language models to embed text . they propose generating diverse transformations of a sentence that preserve its meaning . |
| Outcome: | The proposed method outperforms existing training-free embedding methods by 2.85 points on the sentence semantic text similarity (STS) benchmark. |
KV-Latent: Dimensional-level KV Cache Reduction with Frequency-aware Rotary Positional Embedding (2025.acl-long)
Copied to clipboard
| Challenge: | Large language models (LLMs) based on Transformer Decoders have become the preferred choice for conversational generative AI. |
| Approach: | They propose a paradigm called KV-Latent to reduce the KV cache footprint and improve inference speed by down-sampling the Key-Value vector dimensions into a latent space. |
| Outcome: | The proposed paradigm reduces the KV Cache footprint and improves inference speed with a small amount of extra training, less than 1% of pre-training takes. |
Exploring the Hidden Capacity of LLMs for One-Step Text Generation (2025.emnlp-main)
Copied to clipboard
| Challenge: | Large language models (LLMs) can reconstruct surprisingly long texts via autoregressive generation from just one trained input embedding. |
| Approach: | They show that large language models can reconstruct surprisingly long texts via autoregressive generation from just one trained input embedding. |
| Outcome: | The proposed model can generate hundreds of accurate tokens in one token-parallel forward pass, when provided with only two learned embeddings. |
Token Prepending: A Training-Free Approach for Eliciting Better Sentence Embeddings from LLMs (2025.acl-long)
Copied to clipboard
| Challenge: | Recent studies have focused on prompt engineering to extract sentence embeddings from large language models (LLMs) but these models are mostly decoder-only and the earlier tokens in the sentence cannot attend to the latter, resulting in biased encoding of sentence information and cascading effects on the final decoded token. |
| Approach: | They propose a plug-and-play and training-free technique that prepends each layer’s decoded sentence embedding to the beginning of the sentence in the next layer’ s input. |
| Outcome: | The proposed technique can significantly improve the performance of existing prompt-based sentence embedding methods across different LLMs while incurring negligible additional inference cost. |
Improving Text Embeddings with Large Language Models (2024.acl-long)
Copied to clipboard
| Challenge: | Existing methods for obtaining text embeddings require complex training pipelines . authors leverage proprietary LLMs to generate diverse synthetic data for text embeds based on 93 languages . |
| Approach: | They propose a method for obtaining high-quality text embeddings using only synthetic data and less than 1k training steps. |
| Outcome: | The proposed method achieves strong performance on competitive text embedding benchmarks without using any labeled data. |
LLM-QAT: Data-Free Quantization Aware Training for Large Language Models (2024.findings-acl)
Copied to clipboard
Zechun Liu, Barlas Oguz, Changsheng Zhao, Ernie Chang, Pierre Stock, Yashar Mehdad, Yangyang Shi, Raghuraman Krishnamoorthi, Vikas Chandra
| Challenge: | Several post-training quantization methods have been shown to perform well down to 8-bits. |
| Approach: | They propose a data-free distillation method that leverages generations produced by the pre-trained model to quantize any generative model independent of its training data. |
| Outcome: | The proposed method outperforms SoTA PTQ and LLaMA models at low bit precision. |
T-FREE: Subword Tokenizer-Free Generative LLMs via Sparse Representations for Memory-Efficient Embeddings (2024.emnlp-main)
Copied to clipboard
| Challenge: | Tokenizers are crucial for encoding information in Large Language Models, but their development has stagnated. |
| Approach: | They propose a tokenizer that embeds words through sparse activation patterns over character triplets . they show competitive downstream performance with a parameter reduction of more than 85% . |
| Outcome: | The proposed approach achieves competitive downstream performance with a parameter reduction of more than 85% on embedding layers. |
Tokenizer-Aware Cross-Lingual Adaptation of Decoder-Only LLMs through Embedding Relearning and Swapping (2026.eacl-long)
Copied to clipboard
| Challenge: | Large Language Models (LLMs) have been primarily focused on English, leaving the multilingual ability unexplored. |
| Approach: | They propose a technique that creates new tokenizers and tunes embeddings on fixed model weights for target language adaptation. |
| Outcome: | The proposed method is light-weight and performant but has limitations for older models and high resource languages. |