---
title: "Tencent Runs Text Backwards to Fix a Flaw Hiding in Every LLM"
subtitle: "Researchers from Institute of Science Tokyo and Tencent show that simply reversing input text before embedding boosts frozen LLM performance by up to 8 points on standard benchmarks"
url: "https://alphasignal.ai/news/tencent-runs-text-backwards-to-fix-a-flaw-hiding-in-every-llm"
author: "AlphaSignal Newsroom"
author_url: "https://alphasignal.ai/editorial-team"
company: "AlphaSignal"
company_url: "https://alphasignal.ai/company/alphasignal"
published: "2026-06-04T08:30:49.000Z"
updated: "2026-06-15T00:52:28.558Z"
type: "Paper"
topics:
  - "Llms"
  - "Retrieval"
subtopics:
  - "Long Context"
  - "Embeddings"
upvotes: 12
source_url: "https://arxiv.org/abs/2606.05858"
comments: 0
truncated: true
---

# Tencent Runs Text Backwards to Fix a Flaw Hiding in Every LLM

> Researchers from Institute of Science Tokyo and Tencent show that simply reversing input text before embedding boosts frozen LLM performance by up to 8 points on standard benchmarks

Getting useful text embeddings out of a frozen, off-the-shelf LLM sounds appealing: no fine-tuning, no labeled data, no GPU-hours spent on contrastive training. But there is a structural catch baked into every decoder-only model, and a new paper from researchers at the **Institute of Science Tokyo** and **Tencent** proposes a surprisingly elegant fix: just run the text backwards.

The paper,

_Preview only. Full text requires an AlphaSignal Pro account._

## Sources

- [ReverseEOL: Improving Training-free Text Embeddings via Text Reversal in Decoder-only LLMs](https://arxiv.org/abs/2606.05858)
- [ReverseEOL Full Paper (HTML)](https://arxiv.org/html/2606.05858)
- [KV-Embedding: Training-free Text Embedding via Internal KV Re-routing in Decoder-only LLMs](https://arxiv.org/abs/2601.01046)
