mlsanityData to 14 Oct 2026

Evaluating LLMs Without Oracle Feedback: Agentic Annotation Evaluation Through Unsupervised Consistency Signals

arXiv 2509.08809 ↗First posted 10 Sep 2025cs.CLcited by 0

On the map: LLM agents and tool use → General LLM agent frameworks and benchmarks

Abstract (arXiv)

Large Language Models (LLMs), when paired with prompt-based tasks, have significantly reduced data annotation costs and reliance on human annotators. However, evaluating the quality of their annotations remains challenging in dynamic, unsupervised environments where oracle feedback is scarce and conventional methods fail. To address this challenge, we propose a novel agentic annotation paradigm, where a student model collaborates with a noisy teacher (the LLM) to assess and refine annotation quality without relying on oracle feedback. The student model, acting as an unsupervised feedback mechanism, employs a user preference-based majority voting strategy to evaluate the consistency of the LLM outputs. To systematically measure the reliability of LLM-generated annotations, we introduce the Consistent and Inconsistent (CAI) Ratio, a novel unsupervised evaluation metric. The CAI Ratio not only quantifies the annotation quality of the noisy teacher under limited user preferences but also plays a critical role in model selection, enabling the identification of robust LLMs in dynamic, unsupervised environments. Applied to ten open-domain NLP datasets across four LLMs, the CAI Ratio demonstrates a strong positive correlation with LLM accuracy, establishing it as an essential tool for unsupervised evaluation and model selection in real-world settings.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in General LLM agent frameworks and benchmarks, one of 16 directions in LLM agents and tool use. Builds on 29 of its 36 references that are in the map. First posted on arXiv 10 Sep 2025.

Builds on

29 of 36 references.

References · 29

  1. 2020Language Models are Few-Shot Learnerscited by 32,419
  2. 2022Chain-of-Thought Prompting Elicits Reasoning in Large Language Modelscited by 12,459
  3. 2022ReAct: Synergizing Reasoning and Acting in Language Modelscited by 8,695
  4. 2022Self-Consistency Improves Chain of Thought Reasoning in Language Modelscited by 5,708
  5. 2022Large Language Models are Zero-Shot Reasonerscited by 4,374
  6. 2023Reflexion: Language Agents with Verbal Reinforcement Learningcited by 4,276
  7. 2021Finetuned Language Models Are Zero-Shot Learnerscited by 3,780
  8. 2023Self-Refine: Iterative Refinement with Self-Feedbackcited by 3,442
  9. 2023Tree of Thoughts: Deliberate Problem Solving with Large Language Modelscited by 3,211
  10. 2020MiniLM: Deep Self-Attention Distillation for Task-Agnostic Compression of Pre-Trained Transformerscited by 1,463
  11. 2023Large Language Models Cannot Self-Correct Reasoning Yetcited by 1,017
  12. 2023Can LLMs Express Their Uncertainty? An Empirical Evaluation of Confidence Elicitation in LLMscited by 974
17 more
  1. 2023ChatGPT Outperforms Crowd-Workers for Text-Annotation Taskscited by 870
  2. 2022Large Language Models Can Self-Improvecited by 664
  3. 2020Efficient Intent Detection with Dual Sentence Encoderscited by 433
  4. 2019An Evaluation Dataset for Intent Classification and Out-of-Scope Predictioncited by 386
  5. 2018FewRel: A Large-Scale Supervised Few-Shot Relation Classification Dataset with State-of-the-Art Evaluationcited by 254
  6. 2023Large Language Model Guided Tree-of-Thoughtcited by 221
  7. 2022ZeroGen: Efficient Zero-shot Learning via Dataset Generationcited by 211
  8. 2022Generating Training Data with Language Models: Towards Zero-Shot Language Understandingcited by 199
  9. 2023GraphPrompt: Unifying Pre-Training and Downstream Tasks for Graph Neural Networkscited by 167
  10. 2020MTOP: A Comprehensive Multilingual Task-Oriented Semantic Parsing Benchmarkcited by 158
  11. 2023Active Prompting with Chain-of-Thought for Large Language Modelscited by 154
  12. 2024Relying on the Unreliable: The Impact of Language Models' Reluctance to Express Uncertaintycited by 106
  13. 2023ClusterLLM: Large Language Models as a Guide for Text Clusteringcited by 70
  14. 2020Discovering New Intents with Deep Aligned Clusteringcited by 68
  15. 2024CodecLM: Aligning Language Models with Tailored Synthetic Datacited by 39
  16. 2021TWEAC: Transformer with Extendable QA Agent Classifierscited by 37
  17. 2022New Intent Discovery with Pre-training and Contrastive Learningcited by 28

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2025Can External Validation Tools Improve Annotation Quality for LLM-as-a-Judge?
  2. 2026Auditing Automated Evaluation, Error Propagation, and Runtime Mitigation in Tool-Using Language Agents
  3. 2024Can Large Language Models be Trusted for Evaluation? Scalable Meta-Evaluation of LLMs as Evaluators via Agent Debate
  4. 2025How Humans Help LLMs: Assessing and Incentivizing Human Preference Annotators
  5. 2025Accelerating Unbiased LLM Evaluation via Synthetic Feedback
  6. 2025The Alternative Annotator Test for LLM-as-a-Judge: How to Statistically Justify Replacing Human Annotators with LLMs
  7. 2026BabelJudge: Measuring LLM-as-a-Judge Reliability Across Languages and Agent Trajectories
  8. 2025Unbiased Evaluation of Large Language Models from a Causal Perspective