mlsanityData to 14 Oct 2026

Learning from Emptiness: De-biasing Listwise Rerankers with Content-Agnostic Probability Calibration

arXiv 2604.10150 ↗First posted 11 Apr 2026cs.AI cs.CLcited by 4

Not on the map. Why?

Abstract (arXiv)

Generative listwise reranking leverages global context for superior retrieval but is plagued by intrinsic position bias, where models exhibit structural sensitivity to input order independent of relevance. Existing mitigations present a dilemma: inference-time aggregation incurs prohibitive latency, while training-based methods often fail to eradicate ingrained priors, particularly in compact models. To resolve this dilemma, we propose CapCal (Content-Agnostic Probability Calibration), a training-free framework that mechanically decouples positional bias from ranking decisions. By estimating the bias distribution via content-free placeholders, CapCal rectifies output logits through an entropy-adaptive contrastive mechanism. Evaluations across 10 benchmarks confirm that CapCal achieves superior performance among training-free methods while preserving single-pass efficiency. Notably, it unlocks the latent potential of lightweight models (e.g., 0.6B), delivering absolute NDCG gains exceeding 10 points and outperforming both permutation-based aggregation and data-augmentation baselines.

Key facts. Cited by 4 papers in the map (data through 14 Oct 2026). Builds on 42 of its 53 references that are in the map. First posted on arXiv 11 Apr 2026.

Builds on

42 of 53 references.

References · 42

  1. 2023Lost in the Middle: How Language Models Use Long Contextscited by 3,455
  2. 2016MS MARCO: A Human Generated MAchine Reading COmprehension Datasetcited by 1,605
  3. 2022Contrastive Decoding: Open-ended Text Generation as Optimizationcited by 609
  4. 2023Zephyr: Direct Distillation of LM Alignmentcited by 506
  5. 2023Trusting Your Evidence: Hallucinate Less with Context-aware Decodingcited by 327
  6. 2020Overview of the TREC 2019 deep learning trackcited by 286
  7. 2021Overview of the TREC 2020 deep learning trackcited by 196
  8. 2024Found in the Middle: Calibrating Positional Attention Bias Improves Long Context Utilizationcited by 101
  9. 2025Rank1: Test-Time Compute for Reranking in Information Retrievalcited by 70
  10. 2024FIRST: Faster Improved Listwise Reranking with Single Token Decodingcited by 54
  11. 2024Enhancing Contextual Understanding in Large Language Models through Contrastive Decodingcited by 49
  12. 2024Attention in Large Language Models Yields Efficient Zero-Shot Re-Rankerscited by 43
30 more
  1. 2024ListT5: Listwise Reranking with Fusion-in-Decoder Improves Zero-shot Retrievalcited by 30
  2. 2025Rank-K: Test-Time Reasoning for Listwise Rerankingcited by 29
  3. 2025Overview of the TREC 2022 deep learning trackcited by 22
  4. 2024Exploring User Retrieval Integration towards Large Language Models for Cross-Domain Sequential Recommendationcited by 19
  5. 2024Entropy-Based Decoding for Retrieval-Augmented Large Language Modelscited by 19
  6. 2025RAPID: Efficient Retrieval-Augmented Long Text Generation with Writing Planning and Information Discoverycited by 19
  7. 2025Overview of the TREC 2021 deep learning trackcited by 18
  8. 2023From Classification to Generation: Insights into Crosslingual Retrieval Augmented ICLcited by 17
  9. 2025Thought-Augmented Planning for LLM-Powered Interactive Recommender Agentcited by 17
  10. 2024Distillation Contrastive Decoding: Improving LLMs Reasoning with Contrastive Decoding and Distillationcited by 16
  11. 2024LLM-RankFusion: Mitigating Intrinsic Inconsistency in LLM-based Rankingcited by 16
  12. 2025Generative Large Recommendation Models: Emerging Trends in LLMs for Recommendationcited by 13
  13. 2025Killing Two Birds with One Stone: Unifying Retrieval and Ranking with a Single Generative Recommendation Modelcited by 13
  14. 2025CoSteer: Collaborative Decoding-Time Personalization via Local Delta Steeringcited by 13
  15. 2025FuXi-α: Scaling Recommendation Model with Feature Interaction Enhanced Transformercited by 10
  16. 2024Optimizing Sequential Recommendation Models with Scaling Laws and Approximate Entropycited by 9
  17. 2025Overview of the TREC 2023 deep learning trackcited by 9
  18. 2025SelfAug: Mitigating Catastrophic Forgetting in Retrieval-Augmented Generation via Distribution Self-Alignmentcited by 9
  19. 2026The Next Paradigm Is User-Centric Agent, Not Platform-Centric Servicecited by 9
  20. 2026Why Thinking Hurts: Diagnosing and Rectifying Linguistic Inertia in Large Language Models for Recommendationcited by 9
  21. 2024Inference Scaling for Bridging Retrieval and Augmented Generationcited by 7
  22. 2025Enhancing CTR Prediction with De-correlated Expert Networkscited by 7
  23. 2025A Universal Framework for Compressing Embeddings in CTR Predictioncited by 6
  24. 2026FuXi-Linear: Unleashing the Power of Linear Attention in Long-term Time-aware Sequential Recommendationcited by 6
  25. 2026SpecSteer: Synergizing Local Context and Global Reasoning for Efficient Personalized Generationcited by 6
  26. 2026Can Recommender Systems Teach Themselves? A Recursive Self-Improving Framework with Fidelity Controlcited by 5
  27. 2025Enhancing Input-Label Mapping in In-Context Learning with Contrastive Decodingcited by 4
  28. 2026SPARD: Self-Paced Curriculum for RL Alignment via Integrating Reward Dynamics and Data Utilitycited by 4
  29. 2025DLF: Enhancing Explicit-Implicit Interaction via Dynamic Low-Order-Aware Fusion for CTR Predictioncited by 3
  30. 2026Efficient Personalized Reranking with Semi-Autoregressive Generation and Online Knowledge Distillationcited by 3

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 4.

4 papers.

  1. 2026IE as Cache: Information Extraction Enhanced Agentic Reasoningcited by 3
  2. 2026AsymSpec: Context-Asymmetric Speculative Decoding for Agentic LLMscited by 1
  3. 2026Rethinking the Necessity of Adaptive Retrieval-Augmented Generation through the Lens of Adaptive Listwise Rankingcited by 0
  4. 2026Equal Ranking Quality, Different Decisions: Measuring and Reducing Order Dependence in LLM Scorerscited by 0