NightFeats @ MMU-RAGent NeurIPS 2025: A Context-Optimized Multi-Agent RAG System for the Text-to-Text Track

arXiv cs.CL Papers

Summary

NightFeats is a multi-agent RAG system that won Best Dynamic Evaluation at the MMU-RAGent competition at NeurIPS 2025. It uses a three-phase pipeline (retrieval, curation, composition) and outperforms proprietary baselines on human evaluations.

arXiv:2606.11199v1 Announce Type: new Abstract: We present NightFeats, a structured multi-agent retrieval-augmented generation (RAG) system submitted to the MMU-RAGent competition at NeurIPS 2025, where it was awarded Best Dynamic Evaluation in the text-to-text track. Rather than targeting benchmark maximization, this work proposes a principled pipeline that decomposes knowledge synthesis into three coordinated phases: retrieval, curation, and composition, each governed by explicit intermediate representations and handoff contracts. Inspired by Agentic Context Engineering (ACE), the system introduces temporal-semantic reranking, bounded contradiction reconciliation, and citation-preserving composition as core architectural primitives. Competition results show that NightFeats surpasses proprietary baselines including Claude-SonnetV2 and Nova-Pro on LLM-as-a-Judge and Human Likert evaluations, confirming that architectural transparency and verifiable evidence grounding are better aligned with human preferences than systems optimizing narrowly for automatic similarity metrics.
Original Article
View Cached Full Text

Cached at: 06/11/26, 01:34 PM

# NightFeats @ MMU-RAGent NeurIPS 2025: A Context-Optimized Multi-Agent RAG System for the Text-to-Text Track
Source: [https://arxiv.org/abs/2606.11199](https://arxiv.org/abs/2606.11199)
[View PDF](https://arxiv.org/pdf/2606.11199)

> Abstract:We present NightFeats, a structured multi\-agent retrieval\-augmented generation \(RAG\) system submitted to the MMU\-RAGent competition at NeurIPS 2025, where it was awarded Best Dynamic Evaluation in the text\-to\-text track\. Rather than targeting benchmark maximization, this work proposes a principled pipeline that decomposes knowledge synthesis into three coordinated phases: retrieval, curation, and composition, each governed by explicit intermediate representations and handoff contracts\. Inspired by Agentic Context Engineering \(ACE\), the system introduces temporal\-semantic reranking, bounded contradiction reconciliation, and citation\-preserving composition as core architectural primitives\. Competition results show that NightFeats surpasses proprietary baselines including Claude\-SonnetV2 and Nova\-Pro on LLM\-as\-a\-Judge and Human Likert evaluations, confirming that architectural transparency and verifiable evidence grounding are better aligned with human preferences than systems optimizing narrowly for automatic similarity metrics\.

## Submission history

From: Quentin Fever M \[[view email](https://arxiv.org/show-email/24901e67/2606.11199)\] **\[v1\]**Tue, 21 Apr 2026 19:18:07 UTC \(645 KB\)

Similar Articles

Candidate-Constrained Retrieval-Augmented Generation for LongEval-RAG: System Design and Empirical Analysis

arXiv cs.CL

This paper presents a candidate-constrained RAG system for the LongEval-RAG task at CLEF 2026, combining deterministic provenance tracking with passage retrieval, query expansion, pseudo-relevance feedback, reciprocal rank fusion, evidence reranking, and citation-aware aggregation. An ablation study of ten pipeline variants shows that a rule-based chunking pipeline with sentence-level neural selection achieves the best performance.