mlsanityData to 14 Oct 2026

No=esis: Deterministic-First Retrieval with Two-Tier Context Hydration for Factuality-Critical Queries on Small Local Models

arXiv 2609.07663 ↗First posted 7 Sep 2026cs.IR cs.AI cs.CLcited by 0

Not on the map. Why?

Abstract (arXiv)

A wrong number is worse than no answer. Across factuality-critical domains -- audience metrics, scheduling and rights in media; dosages and lab values in healthcare; figures and citations in finance and legal -- a confident but fabricated value is more damaging than an honest admission of uncertainty. Yet this is the dominant failure mode we observe on small local language models: even when correct evidence is present in context, models fabricate plausible numbers and timestamps. Recent work characterizes a real limit of this regime: below 7B parameters, the bottleneck of retrieval-augmented generation (RAG) is not retrieval quality but context utilization. We present Noesis, the deterministic-first query plane of the Noesis architecture, which makes every deterministic judgment before generation. Its mechanisms follow from the ingestion architecture (subject of a separate patent application): (a) a producer-side fact layer rendering precomputed metric facts verbatim without ranking; (b) positional addressing with deterministic cross-source alignment, resolved ahead of query time at zero LLM cost; (c) provenance scoping as an attribution constraint with multi-tier named-reference routing; and (d) two-tier context with model-triggered verbatim hydration. Across four ablations, a 2B model reaches parity with a 35B model on factual integrity (exact values in all runs; zero confabulated numbers on absent-entity traps); structured retrieval beats flat RAG by +11.4 points at 2B; skeleton-only context preserves quantitative answers at 20-30% smaller prompts; and hydration recovers verbatim narrative in ~8s versus ~29s. Two properties matter for regulated domains: each query resolves in a single generation call, and every reported value is traceable to its exact source and position by construction.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 22 of its 30 references that are in the map. First posted on arXiv 7 Sep 2026.

Builds on

22 of 30 references.

References · 22

  1. 2020Retrieval-Augmented Generation for Knowledge-Intensive NLP Taskscited by 9,297
  2. 2023Lost in the Middle: How Language Models Use Long Contextscited by 3,455
  3. 2023Self-RAG: Learning to Retrieve, Generate, and Critique through Self-Reflectioncited by 1,839
  4. 2024From Local to Global: A Graph RAG Approach to Query-Focused Summarizationcited by 1,453
  5. 2023MemGPT: Towards LLMs as Operating Systemscited by 1,389
  6. 2023RECOMP: Improving Retrieval-Augmented LMs with Compression and Selective Augmentationcited by 256
  7. 2024Corrective Retrieval Augmented Generationcited by 197
  8. 2024RetroLLM: Empowering Large Language Models to Retrieve Fine-grained Evidence within Generationcited by 20
  9. 2026COMI: Coarse-to-fine Context Compression via Marginal Information Gaincited by 13
  10. 2026The Missing Memory Hierarchy: Demand Paging for LLM Context Windowscited by 8
  11. 2025SymRAG: Efficient Neuro-Symbolic Retrieval Through Adaptive Query Routingcited by 5
  12. 2026Retrieval as Generation: A Unified Framework with Self-Triggered Information Planningcited by 5
10 more
  1. 2025CLAUSE: Agentic Neuro-Symbolic Knowledge Graph Reasoning via Dynamic Learnable Context Engineeringcited by 3
  2. 2025ProgRAG: Hallucination-Resistant Progressive Retrieval and Reasoning over Knowledge Graphscited by 3
  3. 2026Can Small Language Models Use What They Retrieve? An Empirical Study of Retrieval Utilization Across Model Scalecited by 3
  4. 2026Structure Guided Retrieval-Augmented Generation for Factual Queriescited by 2
  5. 2026MC-RAG System: A Structure-Driven RAG System for Multi-Constraint Queriescited by 2
  6. 2026Query Symbolically or Retrieve Semantically? A Dataset and Method for Semi-Structured Question Answeringcited by 1
  7. 2026Models for minimalist RAG: B1ade 335M Embedding and 1B Parameter Small Language Modelscited by 1
  8. 2026SeDeM: Selective Decompression of Hidden-State Memories for Long-Context Question Answeringcited by 1
  9. 2026Noesis: Bidirectional Graph-RAG with Adaptive Parallelism and Cross-Knowledge-Base Semantic Discoverycited by 1
  10. 2026TopoCompress: Long Context Compression via Graph-Wired Semantic Trajectoriescited by 1

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2025Worse than Zero-shot? A Fact-Checking Dataset for Evaluating the Robustness of RAG Against Misleading Retrievals
  2. 2024Fact, Fetch, and Reason: A Unified Evaluation of Retrieval-Augmented Generation
  3. 2025RARE: Retrieval-Aware Robustness Evaluation for Retrieval-Augmented Generation Systems
  4. 2026When RAG Fails to Equalize: Geo-bias in Factual Question Answering over Public Companies
  5. 2025Fact or Facsimile? Evaluating the Factual Robustness of Modern Retrievers
  6. 2025Retrieval as a Decision: Training-Free Adaptive Gating for Efficient RAG
  7. 2025Hypothetical Documents or Knowledge Leakage? Rethinking LLM-based Query Expansion
  8. 2025Frustratingly Simple Retrieval Improves Challenging, Reasoning-Intensive Benchmarks