Enhancing Persona Consistency for LLMs’ Role-Playing using Persona-Aware Contrastive Learning (2025.findings-acl)
Copied to clipboard
| Challenge: | Existing methods for analyzing and analyzing large language models (LLMs) lack of emotion and fine-grained role awareness limits the model’s ability to provide personalized and diverse interactions further. |
| Approach: | They propose an annotation-free framework to align LLMs’ behavior during role-playing, enhancing the model’s role consistency. |
| Outcome: | The proposed framework outperforms vanilla LLMs under automatic evaluation methods and human expert evaluation. |
Similar Papers
Aligning Large Language Models via Fully Self-Synthetic Data (2026.acl-long)
Copied to clipboard
| Challenge: | Existing approaches to reinforcement learning from human feedback (RLHF) require expensive human-annotated datasets and proprietary models like GPT-4 to annotate preference pairs. |
| Approach: | They propose a self-synthetic framework for LLM alignment where all training data, including prompts (i.e., user queries), responses, and preferences, are generated by the model itself. |
| Outcome: | The proposed framework enhances the model’s chat capabilities on standard benchmarks like AlpacaEval 2.0 while maintaining strong performance on downstream objective tasks. |
Beyond Static Persona Consistency: Dynamic Persona Coherence in LLM Role-Playing (2026.acl-long)
Copied to clipboard
Yirui QI, Xiaoming Zhang, Ruilin Zeng, Mengyao Liu, Ziyi Zhou, Dezhuang Miao, Bingyu Yan, Zhenyu Guan
| Challenge: | Existing LLMs conflate identity consistency with emotional rigidity . Existing models exhibit either robotic repetition or persona drift . |
| Approach: | They propose a framework that decouples Identity-Layer Stability from Adaptive-Layer Appropriateness to achieve persona coherence repair. |
| Outcome: | Experiments on GPT-4o, Claude-3.5-Sonnet, and DeepSeek-V3.2 show consistent improvements (+16–84% gains) |
CLHA: A Simple Yet Effective Contrastive Learning Framework for Human Alignment (2024.lrec-main)
Copied to clipboard
Feiteng Fang, Liang Zhu, Xi Feng, Jinchang Hou, Qixuan Zhao, Chengming Li, Xiping Hu, Ruifeng Xu, Min Yang
| Challenge: | Large language models (LLMs) have attracted considerable attention from academic and industrial communities due to their outstanding performance in various natural language processing tasks. |
| Approach: | They propose a Contrastive Learning Framework for Human Alignment to evaluate the noise within the data and dynamically adjust the training process. |
| Outcome: | The proposed framework surpasses other algorithms in terms of reward model scores, automatic evaluations, and human assessments on the widely used dataset "Helpful and Harmless" |
Are Economists Always More Introverted? Analyzing Consistency in Persona-Assigned LLMs (2025.findings-emnlp)
Copied to clipboard
| Challenge: | Personalized Large Language Models are increasingly used in diverse applications . prior research examined how well LLMs adhere to predefined personas in writing style . inconsistent responses are influenced by multiple factors, including the assigned persona, stereotypes, and model design choices. |
| Approach: | They propose a standardized framework to analyze consistency in persona-assigned LLMs. |
| Outcome: | The proposed framework evaluates personas across multiple tasks and runs. |
Memory-Driven Role-Playing: Evaluation and Enhancement of Persona Knowledge Utilization in LLMs (2026.findings-acl)
Copied to clipboard
| Challenge: | Existing models fail to recall and accurately apply designated persona knowledge without explicit cues . memory-driven role-playing paradigms are attracting significant interest . |
| Approach: | They propose a memory-driven role-playing paradigm that frames persona knowledge as the LLM's internal memory store and a prompting architecture that guides structured memory retrieval and response generation. |
| Outcome: | The proposed paradigm provides a comprehensive diagnostic for four-stage role-playing abilities across 12 LLMs. |
Aligning LLMs with Individual Preferences via Interaction (2025.coling-main)
Copied to clipboard
| Challenge: | Existing studies on LLMs alignment focus on generalizing their behavior to generalized values such as helpfulness, harmlessness, and honesty. |
| Approach: | They train large language models to "interact to align" to implicitly infer user preferences . they use a multi-turn preference dataset to generate a personalized alignment . |
| Outcome: | The proposed method enables dynamic, personalized alignment via interaction with a multi-turn preference dataset. |
Two Tales of Persona in LLMs: A Survey of Role-Playing and Personalization (2024.findings-emnlp)
Copied to clipboard
| Challenge: | Existing literature on leveraging persona in large language models is disorganized and lacks a systematic taxonomy . leveraging peopleas has resurfaced as an ideal lens for adapting LLMs for specific contexts . |
| Approach: | They propose to categorize current research on leveraging persona in large language models . they propose to use a comprehensive survey to categorize existing studies . |
| Outcome: | The proposed framework is a promising framework for tailoring large language models to specific contexts. |
I Know, but I Don’t Know! How Persona Conflict Undermines Instruction Adherence in Large Language Models (2026.findings-eacl)
Copied to clipboard
| Challenge: | Existing studies on persona-grounded dialogue assume idealized scenarios where persona and user utterances are fully aligned. |
| Approach: | They propose a taxonomy that categorizes model behaviors into three response types . they propose sycophantic, adherent, and wavering responses as response types. |
| Outcome: | The proposed framework categorizes model behaviors into three response types and develops a measurement schema grounded in this taxonomy. |
HER: Human-like Reasoning and Reinforcement Learning for LLM Role-playing (2026.findings-acl)
Copied to clipboard
Chengyu Du, Xintao Wang, Aili Chen, Weiyuan Li, Rui Xu, Junteng Liu, Zishan Huang, Rong Tian, Zijun Sun, Yuhao Li, Liheng Feng, Deming Ding, Pengyu Zhao, Yanghua Xiao
| Challenge: | Existing models for LLM role-playing lack high-quality datasets with explicit reasoning traces and reliable reward signals aligned with human preferences. |
| Approach: | They propose a unified framework for cognitive-level persona simulation that strictly distinguishes characters’ first-person thinking processes from LLMs’ third-person reasoning. |
| Outcome: | The proposed framework outperforms the Qwen3-32B baseline model and achieves a 30.26% and 14.97% performance on the minimax benchmarks. |
Will I Sound Like Me? Improving Persona Consistency in Dialogues through Pragmatic Self-Consciousness (2020.emnlp-main)
Copied to clipboard
| Challenge: | Existing models for improving consistency often train with additional NLI labels or attach trained extra modules to the generative agent. |
| Approach: | They propose to encode personas into dialogue embeddings and a persona-conditioned dialogue dataset to improve persona consistency. |
| Outcome: | The proposed approach can enforce dialogue agents to refrain from contradictions and improve consistency of existing models. |