mlsanityData to 14 Oct 2026

Revealing Potential Biases in LLM-Based Recommender Systems in the Cold Start Setting

arXiv 2508.20401 ↗First posted 28 Aug 2025cs.IR cs.LGcited by 6

Not on the map. Why?

Abstract (arXiv)

Large Language Models (LLMs) are increasingly used for recommendation tasks due to their general-purpose capabilities. While LLMs perform well in rich-context settings, their behavior in cold-start scenarios, where only limited signals such as age, gender, or language are available, raises fairness concerns because they may rely on societal biases encoded during pretraining. We introduce a benchmark specifically designed to evaluate fairness in zero-context recommendation. Our modular pipeline supports configurable recommendation domains and sensitive attributes, enabling systematic and flexible audits of any open-source LLM. Through evaluations of state-of-the-art models (Gemma 3 and Llama 3.2), we uncover consistent biases across recommendation domains (music, movies, and colleges) including gendered and cultural stereotypes. We also reveal a non-linear relationship between model size and fairness, highlighting the need for nuanced analysis.

Key facts. Cited by 6 papers in the map (data through 14 Oct 2026). Builds on 21 of its 30 references that are in the map. First posted on arXiv 28 Aug 2025.

Builds on

21 of 30 references.

References · 21

  1. 2020Language Models are Few-Shot Learnerscited by 32,419
  2. 2020Scaling Laws for Neural Language Modelscited by 6,586
  3. 2023Efficient Memory Management for Large Language Model Serving with PagedAttentioncited by 5,944
  4. 2020RealToxicityPrompts: Evaluating Neural Toxic Degeneration in Language Modelscited by 1,255
  5. 2020StereoSet: Measuring stereotypical bias in pretrained language modelscited by 924
  6. 2021BBQ: A Hand-Built Bias Benchmark for Question Answeringcited by 701
  7. 2022ToxiGen: A Large-Scale Machine-Generated Dataset for Adversarial and Implicit Hate Speech Detectioncited by 422
  8. 2022Recommendation as Language Processing (RLP): A Unified Pretrain, Personalized Prompt & Predict Paradigm (P5)cited by 402
  9. 2023TALLRec: An Effective and Efficient Tuning Framework to Align Large Language Model with Recommendationcited by 337
  10. 2021Towards Understanding and Mitigating Social Biases in Language Modelscited by 296
  11. 2023Is ChatGPT Fair for Recommendation? Evaluating Fairness in Large Language Model Recommendationcited by 114
  12. 2017Beyond Parity: Fairness Objectives for Collaborative Filteringcited by 93
9 more
  1. 2017Multisided Fairness for Recommendationcited by 70
  2. 2023Large Language Models are Competitive Near Cold-start Recommenders for Language- and Item-based Preferencescited by 66
  3. 2024Certifying Counterfactual Bias in LLMscited by 10
  4. 2024Unveiling Bias in Fairness Evaluations of Large Language Models: A Critical Literature Review of Music and Movie Recommendation Systemscited by 7
  5. 2024Unveiling and Mitigating Bias in Large Language Model Recommendations: A Path to Fairnesscited by 6
  6. 2021Ethics and Governance of Artificial Intelligence: Evidence from a Survey of Machine Learning Researcherscited by 5
  7. 2025Text2Playlist: Generating Personalized Playlists from Text on Deezercited by 3
  8. 2025Unmasking Gender Bias in Recommendation Systems and Enhancing Category-Aware Fairnesscited by 3
  9. 2025FilterLLM: Text-To-Distribution LLM for Billion-Scale Cold-Start Recommendationcited by 2

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 6.

6 papers.

  1. 2026Prominence-Stratified Failure Modes in Retrieval-Augmented Commercial Recommendation: A 37,000-Run Auditcited by 1
  2. 2026HELM: A Human-Centered Evaluation Framework for LLM-Powered Recommender Systemscited by 0
  3. 2026Combating data scarcity in recommendation services: Integrating cognitive types of VARK and neural network technologies (LLM)cited by 0
  4. 2026An Investigation of Linguistic Biases in LLM-Based Recommendationscited by 0
  5. 2026Rethinking Fairness in LLM-Based Recommender Systems: A Surveycited by 0
  6. 2026Critical Acclaim Orientation in Large Language Models: Evidence from Film Preference Elicitationcited by 0