mlsanityData to 14 Oct 2026

Teacher-Student Training for Debiasing: General Permutation Debiasing for Large Language Models

arXiv 2403.13590 ↗First posted 20 Mar 2024cs.CLcited by 8

On the map: Knowledge distillation → Knowledge distillation for LLMs

Abstract (arXiv)

Large Language Models (LLMs) have demonstrated impressive zero-shot capabilities and versatility in NLP tasks, however they sometimes fail to maintain crucial invariances for specific tasks. One example is permutation sensitivity, where LLMs' outputs may significantly vary depending on the order of the input options. While debiasing techniques can mitigate these issues, and yield better performance and reliability, they often come with a high computational cost at inference. This paper addresses this inefficiency at inference time. The aim is to distill the capabilities of a computationally intensive, debiased, teacher model into a more compact student model. We explore two variants of student models: one based on pure distillation, and the other on an error-correction approach for more complex tasks, where the student corrects a single biased decision from the teacher to achieve a debiased output. Our approach is general and can be applied to both black-box and white-box LLMs. Furthermore, we demonstrate that our compact, encoder-only student models can outperform their larger, biased teacher counterparts, achieving better results with significantly fewer parameters.

Key facts. Cited by 8 papers in the map (data through 14 Oct 2026). Placed in Knowledge distillation for LLMs, one of 12 directions in Knowledge distillation. Builds on 29 of its 36 references that are in the map. First posted on arXiv 20 Mar 2024.

Builds on

29 of 36 references.

References · 29

  1. 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,399
  2. 2020Language Models are Few-Shot Learnerscited by 32,419
  3. 2023GPT-4 Technical Reportcited by 18,396
  4. 2022Training language models to follow instructions with human feedbackcited by 15,971
  5. 2019Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformercited by 14,599
  6. 2019RoBERTa: A Robustly Optimized BERT Pretraining Approachcited by 13,372
  7. 2023Llama 2: Open Foundation and Fine-Tuned Chat Modelscited by 13,028
  8. 2015Distilling the Knowledge in a Neural Networkcited by 10,324
  9. 2023Judging LLM-as-a-Judge with MT-Bench and Chatbot Arenacited by 8,626
  10. 2017On Calibration of Modern Neural Networkscited by 4,960
  11. 2018Think you have Solved Question Answering? Try ARC, the AI2 Reasoning Challengecited by 4,886
  12. 2021Finetuned Language Models Are Zero-Shot Learnerscited by 3,780
17 more
  1. 2022Scaling Instruction-Finetuned Language Modelscited by 3,024
  2. 2020DeBERTa: Decoding-enhanced BERT with Disentangled Attentioncited by 2,181
  3. 2021Calibrate Before Use: Improving Few-Shot Performance of Language Modelscited by 1,392
  4. 2019TinyBERT: Distilling BERT for Natural Language Understandingcited by 1,167
  5. 2023Large Language Models are not Fair Evaluatorscited by 1,085
  6. 2022Large Language Models Are Human-Level Prompt Engineerscited by 1,045
  7. 2022Super-NaturalInstructions: Generalization via Declarative Instructions on 1600+ NLP Taskscited by 868
  8. 2017RACE: Large-scale ReAding Comprehension Dataset From Examinationscited by 861
  9. 2023Quantifying Language Models' Sensitivity to Spurious Features in Prompt Design or: How I learned to start worrying about prompt formattingcited by 817
  10. 2020SummEval: Re-evaluating Summarization Evaluationcited by 674
  11. 2021Prompt Programming for Large Language Models: Beyond the Few-Shot Paradigmcited by 606
  12. 2023Large Language Models Are Not Robust Multiple Choice Selectorscited by 486
  13. 2019Cosmos QA: Machine Reading Comprehension with Contextual Commonsense Reasoningcited by 285
  14. 2023Large Language Models Sensitivity to The Order of Options in Multiple-Choice Questionscited by 263
  15. 2022Leveraging Large Language Models for Multiple Choice Question Answeringcited by 203
  16. 2023The Larger They Are, the Harder They Fail: Language Models do not Recognize Identifier Swaps in Pythoncited by 29
  17. 2023Mitigating Word Bias in Zero-shot Prompt-based Classifierscited by 9

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2024: 2; 2025: 3; 2026: 3.

8 papers.

  1. 2024Efficient LLM Comparative Assessment: a Product of Experts Framework for Pairwise Comparisonscited by 22
  2. 2024Mitigating Selection Bias with Node Pruning and Auxiliary Optionscited by 12
  3. 2026Who can we trust? LLM-as-a-jury for Comparative Assessmentcited by 7
  4. 2025Hearing the Order: Investigating Position Bias in Large Audio-Language Modelscited by 5
  5. 2025Generalised Probabilistic Modelling and Improved Uncertainty Estimation in Comparative LLM-as-a-judgecited by 4
  6. 2025SCOPE: Stochastic and Counterbiased Option Placement for Evaluating Large Language Modelscited by 2
  7. 2026Are LLMs Positionally Consistent Ordinal Classifiers? A Systematic Evaluationcited by 1
  8. 2026Models Know Models Best: Evaluation via Model-Preferred Formatscited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025Do Students Debias Like Teachers? On the Distillability of Bias Mitigation Methods
  2. 2026Mutual Debiasing via Dual-Seed Comparison for Probabilistic Sampling in Large Language Models
  3. 2024A Probabilistic Perspective on Unlearning and Alignment for Large Language Models
  4. 2024Towards Resource Efficient and Interpretable Bias Mitigation in Large Language Models
  5. 2026Debias-SparseGPT: Bias-Aware Pruning for Large Language Models
  6. 2023Deduction under Perturbed Evidence: Probing Student Simulation Capabilities of Large Language Models
  7. 2024A Multi-LLM Debiasing Framework
  8. 2025BiasFilter: An Inference-Time Debiasing Framework for Large Language Models