mlsanityData to 14 Oct 2026

Bayesian Test-Time Inference of Task-Aligned Similarity from Weak Interactive Feedback

arXiv 2610.18656 ↗First posted 13 Oct 2026cs.LG cs.CLcited by 0

Not on the map. Why?

Abstract (arXiv)

Language model systems repeatedly make test-time decisions by comparing similarity: they retrieve exemplars, rank candidates, and revise responses. Yet similarity is not unique: lexical, semantic, structural, and stylistic resemblance can point to different candidates, and the task-aligned notion can change across instances. We introduce BASIS (Bayesian Axis Selection via Interactive Signals), a training-free Bayesian method that uses weak test-time feedback to infer which criterion in a supplied, task-specific library governs the current decision. It maintains a joint posterior over candidate targets and criteria, updates it from scalar scores, pairwise preferences, or verifier outputs, and selects probes by expected information gain. In a finite noisy model, we characterize when the target-criterion pair is recoverable and show why fixed-criterion and fixed-mixture rules can fail under criterion shift. Across concept recovery, reasoning exemplar selection, and response refinement, BASIS improves six-probe concept-recovery Success@1 from 56.0% to 68.1%, reasoning accuracy from 72.3% to 73.1%, and refinement win rate from 61.4% to 63.8% over tuned static mixtures. It also exceeds compute-matched best-of-N selection on reasoning (72.1%) and refinement (61.2%). Gains are largest under criterion mismatch and with larger candidate pools, and matched controls indicate that adaptive criterion weighting explains much of the benefit, with joint target-criterion inference adding further value. BASIS selects only among supplied criteria and does not discover new ones.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 22 of its 31 references that are in the map. First posted on arXiv 13 Oct 2026.

Builds on

22 of 31 references.

References · 22

  1. 2020Language Models are Few-Shot Learnerscited by 32,419
  2. 2021Training Verifiers to Solve Math Word Problemscited by 10,109
  3. 2020Retrieval-Augmented Generation for Knowledge-Intensive NLP Taskscited by 9,297
  4. 2023Reflexion: Language Agents with Verbal Reinforcement Learningcited by 4,276
  5. 2022Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedbackcited by 3,720
  6. 2023Self-Refine: Iterative Refinement with Self-Feedbackcited by 3,442
  7. 2023Self-RAG: Learning to Retrieve, Generate, and Critique through Self-Reflectioncited by 1,839
  8. 2022Challenging BIG-Bench Tasks and Whether Chain-of-Thought Can Solve Themcited by 1,807
  9. 2016ConceptNet 5.5: An Open Multilingual Graph of General Knowledgecited by 1,067
  10. 2024Large Language Monkeys: Scaling Inference Compute with Repeated Samplingcited by 991
  11. 2023CRITIC: Large Language Models Can Self-Correct with Tool-Interactive Critiquingcited by 783
  12. 2021Did Aristotle Use a Laptop? A Question Answering Benchmark with Implicit Reasoning Strategiescited by 743
10 more
  1. 2023OpenAssistant Conversations -- Democratizing Large Language Model Alignmentcited by 721
  2. 2021Learning To Retrieve Prompts for In-Context Learningcited by 648
  3. 2024Rewarding Progress: Scaling Automated Process Verifiers for LLM Reasoningcited by 279
  4. 2014SimLex-999: Evaluating Semantic Models with (Genuine) Similarity Estimationcited by 172
  5. 2023In-Context Learning through the Bayesian Prismcited by 83
  6. 2024FollowIR: Evaluating and Teaching Information Retrieval Models to Follow Instructionscited by 69
  7. 2024RaFe: Ranking Feedback Improves Query Rewriting for RAGcited by 49
  8. 2022Interpretable by Design: Learning Predictors by Composing Interpretable Queriescited by 28
  9. 2024Bayesian Example Selection Improves In-Context Learning for Speech, Text, and Visual Modalitiescited by 8
  10. 2024Reasoning Graph Enhanced Exemplars Retrieval for In-Context Learningcited by 3

Built on by

Not cited yet.