Noēsis: Deterministic-First Retrieval with Two-Tier Context Hydration for Factuality-Critical Queries on Small Local Models

• URL: http://arxiv.org/abs/2609.07663v1
• Date: Mon, 07 Sep 2026 15:49:16 GMT
• Title: Noēsis: Deterministic-First Retrieval with Two-Tier Context Hydration for Factuality-Critical Queries on Small Local Models
• Abstract summary: We present Noesis, the deterministic-first query plane of the Noesis architecture.<n>Noesis makes every deterministic judgment before generation.<n>Every reported value is traceable to its exact source and position by construction.
• Score: 0.0
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: 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.

Related papers

• SynthSentry: Detecting Synthetic Data Contamination in Language Model Training Data [2.998247765202193]
  We introduce a corpus-level, model-agnostic contamination signal requiring no access to the generating model, no generation history, and no synthetic labels.<n>We evaluate on corpora contaminated by small open-weight generators and an instruction-tuned open-weight model under a leave-one-generator-out protocol.<n>The score ranks corpora by severity with little loss when whole generator families are held out.
  arXiv  Detail & Related papers  (2026-09-11T02:25:18Z)
• Never the Number: Structural Abstention for AI Systems Whose Answers Are Consumed as Fact [0.0]
  Large language models have made natural language interfaces to databases newly credible, but calibrated text-to-aggregated systems fail in a way that matters for deployment.<n>We propose an architectural pattern for such systems, a trusted kernel with a generative shell, resting on one invariant.<n>A generative shell interprets under input and phrases replies; a deterministic kernel matches fully specified questions against a bounded set of answerable question shapes.<n>The two meet at a confirmation the user reads before any value is computed, and requests the kernel cannot express are declined rather than approximated.
  arXiv  Detail & Related papers  (2026-08-14T03:53:21Z)
• DiagLoop: A Counterfactual Data Flywheel with Stage-Localized Reinforcement for Diagnostic LLMs [5.241146164724621]
  We present DiagLoop, a counterfactual data flywheel that converts codified physical relations or clinical guidelines into training supervision.<n>A training-only teacher proposes counterfactual worlds by varying causes, contexts, and observations, while an independent hybrid checker admits only valid worlds.<n>For nonterminal failures, a bounded repair probes downstream competence, and the resulting weakness profile guides subsequent data generation.
  arXiv  Detail & Related papers  (2026-08-04T13:49:41Z)
• Two-Level Meta-Rubrics for Evaluating Open-Ended Generation: GAMUT, a Benchmark for Factual Completeness [8.98920719663984]
  Gamut is a benchmark for factual completeness in long-form generation.<n>We construct 1,813 questions grounded in real wearable imagery across 10 diverse domains.
  arXiv  Detail & Related papers  (2026-07-21T17:42:50Z)
• Before the Action: Benchmarking LLMs on Prospective Hypothesis Discovery [95.44432473367836]
  Large language models (LLMs) excel at answering pre-specified questions, yet their ability to navigate the open-ended, pre-conclusion stage of discovery remains largely unmeasured.<n>We introduce Prospective Hypothesis Discovery (PHD), which asks models to autonomously construct grounded, discnative, and testable hypothesis spaces from inconclusive evidence.<n>We introduce HypoArena, comprising HypoData, a benchmark of 988 cases across six scientific and analytical domains, and HypoEval, an evaluation framework for open-ended hypothesis sets.
  arXiv  Detail & Related papers  (2026-07-17T08:56:43Z)
• Resolving Ambiguity in Composed Image Retrieval via Calibrated Interaction [0.0]
  Composed image retrieval (CIR) searches a corpus with a reference image and a text describing how to modify it.<n>We introduce AmbiCIR, a benchmark and human-validated user simulator that revives the dormant auxiliary and dialogue annotations of CIRR.
  arXiv  Detail & Related papers  (2026-05-23T15:49:16Z)
• WESR: Scaling and Evaluating Word-level Event-Speech Recognition [59.21814194620928]
  Speech conveys not only linguistic information but also rich non-verbal vocal events such as laughing and crying.<n>We develop a refined taxonomy of 21 vocal events, with a new categorization into discrete (standalone) versus continuous (mixed with speech) types.<n>Based on the refined taxonomy, we introduce WESR-Bench, an expert-annotated evaluation set (900+ utterances) with a novel position-aware protocol.
  arXiv  Detail & Related papers  (2026-01-08T02:23:21Z)
• On Reference (In-)Determinacy in Natural Language Inference [62.904689974282334]
  We revisit the reference determinacy (RD) assumption in the task of natural language inference (NLI)<n>We observe that current NLI models fail in downstream applications such as fact verification, where the input premise and hypothesis may refer to different contexts.<n>We introduce RefNLI, a diagnostic benchmark for identifying reference ambiguity in NLI examples.
  arXiv  Detail & Related papers  (2025-02-09T06:58:13Z)
• Fine-tuning Language Models for Factuality [96.5203774943198]
  Large pre-trained language models (LLMs) have led to their widespread use, sometimes even as a replacement for traditional search engines. Yet language models are prone to making convincing but factually inaccurate claims, often referred to as 'hallucinations' In this work, we fine-tune language models to be more factual, without human labeling.
  arXiv  Detail & Related papers  (2023-11-14T18:59:15Z)
• Generating Benchmarks for Factuality Evaluation of Language Models [61.69950787311278]
  We propose FACTOR: Factual Assessment via Corpus TransfORmation, a scalable approach for evaluating LM factuality. FACTOR automatically transforms a factual corpus of interest into a benchmark evaluating an LM's propensity to generate true facts from the corpus vs. similar but incorrect statements. We show that: (i) our benchmark scores increase with model size and improve when the LM is augmented with retrieval; (ii) benchmark score and perplexity do not always agree on model ranking; (iii) when perplexity and benchmark score disagree, the latter better reflects factuality in open-ended generation.
  arXiv  Detail & Related papers  (2023-07-13T17:14:38Z)
• WiCE: Real-World Entailment for Claims in Wikipedia [63.234352061821625]
  We propose WiCE, a new fine-grained textual entailment dataset built on natural claim and evidence pairs extracted from Wikipedia. In addition to standard claim-level entailment, WiCE provides entailment judgments over sub-sentence units of the claim. We show that real claims in our dataset involve challenging verification and retrieval problems that existing models fail to address.
  arXiv  Detail & Related papers  (2023-03-02T17:45:32Z)
• Factual Error Correction for Abstractive Summaries Using Entity Retrieval [57.01193722520597]
  We propose an efficient factual error correction system RFEC based on entities retrieval post-editing process. RFEC retrieves the evidence sentences from the original document by comparing the sentences with the target summary. Next, RFEC detects the entity-level errors in the summaries by considering the evidence sentences and substitutes the wrong entities with the accurate entities from the evidence sentences.
  arXiv  Detail & Related papers  (2022-04-18T11:35:02Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.