mlsanityData to 7 Oct 2026

Enhancing Temporal Sensitivity of Large Language Model for Recommendation with Counterfactual Tuning

arXiv 2507.03047 ↗First posted 3 Jul 2025cs.CL cs.AI cs.IRcited by 0

On the map: LLM reasoning → LLM reasoning for recommendation

Abstract (arXiv)

Recent advances have applied large language models (LLMs) to sequential recommendation, leveraging their pre-training knowledge and reasoning capabilities to provide more personalized user experiences. However, existing LLM-based methods fail to sufficiently leverage the rich temporal information inherent in users' historical interaction sequences, stemming from fundamental architectural constraints: LLMs process information through self-attention mechanisms that lack inherent sequence ordering and rely on position embeddings designed primarily for natural language rather than user interaction sequences. This limitation significantly impairs their ability to capture the evolution of user preferences over time and predict future interests accurately. To address this critical gap, we propose underlineCounterfactual underlineEnhanced underlineTemporal Framework for LLM-Based underlineRecommendation (CETRec). CETRec is grounded in causal inference principles, which allow it to isolate and measure the specific impact of temporal information on recommendation outcomes. Combined with our counterfactual tuning task derived from causal analysis, CETRec effectively enhances LLMs' awareness of both absolute order (how recently items were interacted with) and relative order (the sequential relationships between items). Extensive experiments on real-world datasets demonstrate the effectiveness of our CETRec. Our code is available at https://anonymous.4open.science/r/CETRec-B9CE/.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Placed in LLM reasoning for recommendation, one of 17 directions in LLM reasoning. Builds on 29 of its 58 references that are in the map. First posted on arXiv 3 Jul 2025.

Builds on

29 of 58 references.

References · 29

  1. 2017Attention Is All You Needcited by 59,932
  2. 2021RoFormer: Enhanced Transformer with Rotary Position Embeddingcited by 4,581
  3. 2018Self-Attentive Sequential Recommendationcited by 1,169
  4. 2017Counterfactual Explanations without Opening the Black Box: Automated Decisions and the GDPRcited by 1,086
  5. 2015Session-based Recommendations with Recurrent Neural Networkscited by 685
  6. 2018Personalized Top-N Sequential Recommendation via Convolutional Sequence Embeddingcited by 473
  7. 2023A Survey on Large Language Models for Recommendationcited by 378
  8. 2023TALLRec: An Effective and Efficient Tuning Framework to Align Large Language Model with Recommendationcited by 336
  9. 2023LLMRec: Large Language Models with Graph Augmentation for Recommendationcited by 171
  10. 2023Adapting Large Language Models by Integrating Collaborative Semantics for Recommendationcited by 171
  11. 2023Uncovering ChatGPT's Capabilities in Recommender Systemscited by 160
  12. 2018A Simple Convolutional Generative Network for Next Item Recommendationcited by 120
17 more
  1. 2023InferFix: End-to-End Program Repair with LLMscited by 113
  2. 2024Learnable Item Tokenization for Generative Recommendationcited by 106
  3. 2021Position Information in Transformers: An Overviewcited by 103
  4. 2023How to Index Item IDs for Recommendation Foundation Modelscited by 102
  5. 2023A Bi-Step Grounding Paradigm for Large Language Models in Recommendation Systemscited by 73
  6. 2024IDGenRec: LLM-RecSys Alignment with Textual ID Learningcited by 50
  7. 2020Clicks can be Cheating: Counterfactual Recommendation for Mitigating Clickbait Issuecited by 48
  8. 2021CauseRec: Counterfactual User Sequence Synthesis for Sequential Recommendationcited by 46
  9. 2023E4SRec: An Elegant Effective Efficient Extensible Solution of Large Language Models for Sequential Recommendationcited by 39
  10. 2019CosRec: 2D Convolutional Neural Networks for Sequential Recommendationcited by 37
  11. 2023Bridging Items and Language: A Transition Paradigm for Large Language Model-Based Recommendationcited by 29
  12. 2021Achieving Counterfactual Fairness for Causal Banditcited by 15
  13. 2023Temporal Interest Network for User Response Predictioncited by 9
  14. 2024Improve Temporal Awareness of LLMs for Sequential Recommendationcited by 8
  15. 2024Causality-Enhanced Behavior Sequence Modeling in LLMs for Personalized Recommendationcited by 8
  16. 2023Path-Specific Counterfactual Fairness for Recommender Systemscited by 7
  17. 2024GPT4Rec: Graph Prompt Tuning for Streaming Recommendationcited by 7

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026CALMRec: Causally Aligned Language Memory for Long-Horizon Recommendation
  2. 2025What Makes LLMs Effective Sequential Recommenders? A Study on Preference Intensity and Temporal Context
  3. 2024LLM-ESR: Large Language Models Enhancement for Long-tailed Sequential Recommendation
  4. 2025User Feedback Alignment for LLM-powered Exploration in Large-scale Recommendation Systems
  5. 2024Rethinking Large Language Model Architectures for Sequential Recommendations
  6. 2025EAGER-LLM: Enhancing Large Language Models as Recommenders through Exogenous Behavior-Semantic Integration
  7. 2026Principled Synthetic Data Enables the First Scaling Laws for LLMs in Recommendation
  8. 2024LLM4DSR: Leveraging Large Language Model for Denoising Sequential Recommendation