# DecisionRank: On-Policy Distillation for Decision-Model Rerankers

Source: [arXiv](https://arxiv.org/abs/2610.42128v1)  
Feed7 permalink: https://feed7.dev/p/2610-42128v1-0sepil9  
Published: 2026-10-30T17:56:58.000Z  
Trust: Needs Review (needs_review)

## Why Included

DecisionRank transfers a generative ranker's preferences into a score-based reranker through on-policy distillation. The paper reports competitive retrieval results with substantially faster ranking at 8B parameters.

## Source Summary

DecisionRank sorts a decision model's document scores into a ranking, then uses a generative teacher's conditional feedback to supervise those scores. The paper reports that **1.7B** DecisionRank beats its **27B** teacher on TREC DL19–23; its 8B model runs **24.5×** faster on TREC DL and **16.7×** faster on BRIGHT than a generative listwise counterpart.

## Practical Implication

Builders using retrieval in coding agents could evaluate a score-based reranker when ranking latency matters. The method combines item-space on-policy distillation with offline ranking supervision, and the paper reports results on TREC DL, BEIR and BRIGHT.

## Agent-Ready Context

DecisionRank sorts a decision model's document scores into a ranking, then uses a generative teacher's conditional feedback to supervise those scores. The paper reports that **1.7B** DecisionRank beats its **27B** teacher on TREC DL19–23; its 8B model runs **24.5×** faster on TREC DL and **16.7×** faster on BRIGHT than a generative listwise counterpart.

Builders using retrieval in coding agents could evaluate a score-based reranker when ranking latency matters. The method combines item-space on-policy distillation with offline ranking supervision, and the paper reports results on TREC DL, BEIR and BRIGHT.

These are benchmark results, not evidence about end-to-end agent quality or latency in a specific retrieval stack. The speed comparisons apply to the named benchmarks and generative listwise counterpart.

## Connected Context

Feed7 judgment across 974 accumulated Signals:

DecisionRank offers a fast score-based reranker trained from generative feedback, giving retrieval systems a way to reduce ranking latency without replacing their candidate-generation stage. DecisionRAG explores the opposite architectural move by navigating a corpus without a first-stage retriever. The benchmark results motivate comparing these routes, but do not establish which improves an end-to-end agent.

- [Retriever-Free Retrieval: Navigating Documents with Decision Models](https://feed7.dev/p/2610-35868v1-1uzxrmk) — DecisionRank speeds ranking after candidates are available; DecisionRAG removes the first-stage retriever and navigates document branches, so they change different stages of retrieval.
- [Yalda-Embedding: Open Development of a Competitive Persian Retrieval Model](https://feed7.dev/p/2610-32933v1-0kbcu6k) — Yalda provides a compact retrieval model, while DecisionRank focuses on reranking. They could serve complementary stages, though the supplied evidence does not evaluate them together.

## Context Map

- Layer: model
- Domains: research, data
- Topics: retrieval

## Uncertainty

- These are benchmark results, not evidence about end-to-end agent quality or latency in a specific retrieval stack. The speed comparisons apply to the named benchmarks and generative listwise counterpart.

## Agent Instruction

Use this item as source-backed context. Do not invent claims beyond the linked source. If this item conflicts with another source, call out the conflict.