Does LLM Focus on the Right Words? Diagnosing Language Bias in LLM-based Recommenders

• URL: http://arxiv.org/abs/2510.10978v1
• Date: Mon, 13 Oct 2025 03:35:26 GMT
• Title: Does LLM Focus on the Right Words? Diagnosing Language Bias in LLM-based Recommenders
• Abstract summary: Group Distributionally Robust Optimization-based Tuning (GDRT) is a novel fine-tuning paradigm that enforces consistent model performance across token groups with varying degrees of relevance to auxiliary tokens.<n>Extensive experiments conducted on three public datasets demonstrate that GDRT effectively mitigates language bias, yielding substantial improvements in recommendation accuracy.
• Score: 19.30200818581719
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Large language models (LLMs), owing to their extensive open-domain knowledge and semantic reasoning capabilities, have been increasingly integrated into recommender systems (RS). However, a substantial gap remains between the pre-training objectives of LLMs and the specific requirements of recommendation tasks. To address this gap, supervised fine-tuning (SFT) is commonly performed on specially curated recommendation datasets to further enhance their predictive ability. Despite its success, SFT exhibits a critical limitation: it induces Language Bias, whereby the model over-relies on auxiliary tokens-such as task descriptions and prefix-generated tokens-while underutilizing core user interaction tokens that encode user-specific preferences. This bias not only undermines recommendation accuracy but also raises unfairness concerns. To address this issue, we propose Group Distributionally Robust Optimization-based Tuning (GDRT), a novel fine-tuning paradigm that enforces consistent model performance across token groups with varying degrees of relevance to auxiliary tokens. By adaptively upweighting underperforming groups, typically those weakly correlated with auxiliary tokens, GDRT shifts the model's attention from superficial auxiliary cues to informative user interaction tokens, thereby mitigating language bias. Extensive experiments conducted on three public datasets demonstrate that GDRT effectively mitigates language bias, yielding substantial improvements in recommendation accuracy (with an average NDCG@10 gain of 24.29%) and significantly enhancing recommendation fairness.

Related papers

• Preference Data Selection for Mitigating the Alignment Tax in Large Language Models [52.18310687523229]
  BALIGN is a balanced data selection strategy that mitigates catastrophic forgetting while optimizing alignment efficacy.<n>By aggregating three key data-centric features into a unified composite risk score, BALIGN systematically filters out high-risk preference samples.<n>Experiments on standard human preference datasets demonstrate that BALIGN strongly preserves foundational capabilities without compromising alignment gains.
  arXiv  Detail & Related papers  (2026-08-25T07:58:42Z)
• MAESTRO: Meta-learning Adaptive Estimation of Scalarization Trade-offs for Reward Optimization [56.074760766965085]
  Group-Relative Policy Optimization has emerged as an efficient paradigm for aligning Large Language Models (LLMs)<n>We propose MAESTRO, which treats reward scalarization as a dynamic latent policy, leveraging the model's terminal hidden states as a semantic bottleneck.<n>We formulate this as a contextual bandit problem within a bi-level optimization framework, where a lightweight Conductor network co-evolves with the policy by utilizing group-relative advantages as a meta-reward signal.
  arXiv  Detail & Related papers  (2026-01-12T05:02:48Z)
• Multimodal Large Language Models with Adaptive Preference Optimization for Sequential Recommendation [60.33386541343322]
  We propose a Multimodal Large Language Models framework that integrates Hardness-aware and Noise-regularized preference optimization for Recommendation (HaNoRec)<n>Specifically, HaNoRec dynamically adjusts optimization weights based on both the estimated hardness of each training sample and the policy model's real-time responsiveness.
  arXiv  Detail & Related papers  (2025-11-24T04:10:46Z)
• Distribution Preference Optimization: A Fine-grained Perspective for LLM Unlearning [26.120338506874976]
  Unlearning, which aims to remove the influence of specific data while preserving overall model utility, is becoming an important research area.<n>We derive a novel unlearning algorithm termed textbfDistribution textbfPreference textbfOptimization (DiPO)<n>DiPO attains the highest forget quality on the TOFU benchmark, and maintains leading scalability and sustainability on the MUSE benchmark.
  arXiv  Detail & Related papers  (2025-10-06T12:49:00Z)
• IGD: Token Decisiveness Modeling via Information Gain in LLMs for Personalized Recommendation [79.22388408461458]
  We introduce an Information Gain-based Decisiveness-aware Token handling (IGD) strategy that integrates token decisiveness into both tuning and decoding.<n>IGD consistently improves recommendation accuracy, achieving significant gains on widely used ranking metrics compared to strong baselines.
  arXiv  Detail & Related papers  (2025-06-16T08:28:19Z)
• A Novel Generative Model with Causality Constraint for Mitigating Biases in Recommender Systems [20.672668625179526]
  Latent confounding bias can obscure the true causal relationship between user feedback and item exposure.<n>We propose a novel generative framework called Latent Causality Constraints for Debiasing representation learning in Recommender Systems.
  arXiv  Detail & Related papers  (2025-05-22T14:09:39Z)
• Supervised Optimism Correction: Be Confident When LLMs Are Sure [91.7459076316849]
  We establish a novel theoretical connection between supervised fine-tuning and offline reinforcement learning.<n>We show that the widely used beam search method suffers from unacceptable over-optimism.<n>We propose Supervised Optimism Correction, which introduces a simple yet effective auxiliary loss for token-level $Q$-value estimations.
  arXiv  Detail & Related papers  (2025-04-10T07:50:03Z)
• Training Large Recommendation Models via Graph-Language Token Alignment [53.3142545812349]
  We propose a novel framework to train Large Recommendation models via Graph-Language Token Alignment.<n>By aligning item and user nodes from the interaction graph with pretrained LLM tokens, GLTA effectively leverages the reasoning abilities of LLMs.<n> Furthermore, we introduce Graph-Language Logits Matching (GLLM) to optimize token alignment for end-to-end item prediction.
  arXiv  Detail & Related papers  (2025-02-26T02:19:10Z)
• SPRec: Self-Play to Debias LLM-based Recommendation [23.875509546540904]
  Large language models (LLMs) have attracted significant attention in recommendation systems.<n>We propose SPRec, a novel self-play framework designed to mitigate over-recommendation and improve fairness without requiring additional data or manual intervention.
  arXiv  Detail & Related papers  (2024-12-12T12:53:30Z)
• SELF: Surrogate-light Feature Selection with Large Language Models in Deep Recommender Systems [51.09233156090496]
  SurrogatE-Light Feature selection method for deep recommender systems.<n> SELF integrates semantic reasoning from Large Language Models with task-specific learning from surrogate models.<n> Comprehensive experiments on three public datasets from real-world recommender platforms validate the effectiveness of SELF.
  arXiv  Detail & Related papers  (2024-12-11T16:28:18Z)
• RAZOR: Sharpening Knowledge by Cutting Bias with Unsupervised Text Rewriting [16.633948320306832]
  biases prevalent in manually constructed datasets can introduce spurious correlations between tokens and labels.<n>Existing debiasing methods often rely on prior knowledge of specific dataset biases.<n>We propose RAZOR, a novel, unsupervised, and data-focused debiasing approach based on text rewriting for shortcut mitigation.
  arXiv  Detail & Related papers  (2024-12-10T17:02:58Z)
• RosePO: Aligning LLM-based Recommenders with Human Values [38.029251417802044]
  We propose a general framework -- Recommendation with smoothing personalized Preference Optimization (RosePO) RosePO better aligns with customized human values during the post-training stage. Evaluation on three real-world datasets demonstrates the effectiveness of our method.
  arXiv  Detail & Related papers  (2024-10-16T12:54:34Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.