Cross-Context Review: Improving LLM Output Quality by Separating Production and Review Sessions

TL;DR

This paper proposes Cross-Context Review (CCR), a simple method where error review is conducted in a new session without prior context, significantly improving error detection in large language models compared to traditional review methods.

Contribution

The paper introduces CCR, a novel review approach that separates production and review sessions, demonstrating its effectiveness over existing review strategies in error detection tasks.

Findings

CCR outperforms other review methods in error detection accuracy.

Repetition within the same session does not improve review quality.

Context separation is the key factor for CCR's success.

Abstract

Large language models struggle to catch errors in their own outputs when the review happens in the same session that produced them. This paper introduces Cross-Context Review (CCR), a straightforward method where the review is conducted in a fresh session with no access to the production conversation history. We ran a controlled experiment: 30 artifacts (code, technical documents, presentation scripts) with 150 injected errors, tested under four review conditions -- same-session Self-Review (SR), repeated Self-Review (SR2), context-aware Subagent Review (SA), and Cross-Context Review (CCR). Over 360 reviews, CCR reached an F1 of 28.6%, outperforming SR (24.6%, p=0.008, d=0.52), SR2 (21.7%, p<0.001, d=0.72), and SA (23.8%, p=0.004, d=0.57). The SR2 result matters most for interpretation: reviewing twice in the same session did not beat reviewing once (p=0.11), which rules out repetition as an explanation for CCR's advantage. The benefit comes from context separation itself. CCR works with any model, needs no infrastructure, and costs only one extra session.