One Model Is Not a Crowd: Multi-LLM and Aspect-Conditioned Diverse Comment Generation
cs.CL
Submitted: 2026-09-27
Updated: 2026-09-27
License: http://creativecommons.org/licenses/by/4.0/
The gist: Human communication on the internet is shaped by diverse perspectives, most visibly expressed in online comment spaces.
Terminology
Abstract
Human communication on the internet is shaped by diverse perspectives, most visibly expressed in online comment spaces. As large language model (LLM)based AI agents begin to inhabit these spaces, a key question arises: whether synthetic comment threads can capture the diversity inherent in human discourse. This concern is increasingly important, as the growing presence of homogenized AI-generated content risks reducing diversity over time, potentially leading to model collapse and degrading the richness of digital communication. Inspired by the plurality of human crowds and the aspect-driven nature of discourse, we hypothesize that comment diversity is better approximated by combining multiple LLMs with aspect-conditioned generation. We formalize and evaluate this approach using models from different providers and introduce a framework that characterizes diversity across semantic, linguistic, and socio-pragmatic features along three axes: dispersion, coverage, and alignment. Using this framework, we conduct a large-scale study on over 2 million YouTube comments across multiple domains. Our results reveal that multi-LLM and aspect-conditioned generation better align with human comment distributions and such data remains viable under pretraining style curation and is effective for downstream tasks. Yet, human diversity remains unmatched. Overall, our findings provide a practical foundation for generating more diverse and socially grounded discourse in AI-mediated environments.
Sources
- Phi-4 Technical Report
- Phi-4-Mini Technical Report: Compact yet Powerful Multimodal Language Models via Mixture-of-LoRAs
- GPT-4 Technical Report
- The Llama 4 Herd: Architecture, Training, Evaluation, and Deployment Notes
- gpt-oss-120b & gpt-oss-20b Model Card
- Using Semantic Similarity and Text Embedding to Measure the Social Media Echo of Strategic Communications
- Language-Driven Opinion Dynamics in Agent-Based Simulations with LLMs
- On the Diversity of Synthetic Data and its Impact on Training Large Language Models
- From Persona to Personalization: A Survey on Role-Playing Language Agents
- Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilities
- When One LLM Drools, Multi-LLM Collaboration Rules
- Scaling Synthetic Data Creation with 1,000,000,000 Personas
- RealToxicityPrompts: Evaluating Neural Toxic Degeneration in Language Models
- GneissWeb: Preparing High Quality Data for LLMs at Scale
- BERTopic: Neural topic modeling with a class-based TF-IDF procedure
- Training Compute-Optimal Large Language Models
- The Anatomy of the Moltbook Social Graph
- Explaining Length Bias in LLM-Based Preference Evaluations
- Mixtral of Experts
- DeepSeek-V3 Technical Report
Related papers
- Exploring Solution Divergence and Its Effect on Large Language Model Problem Solving
- Ishigaki-IDS-Bench: A Benchmark for Generating Information Delivery Specification from BIM Information Requirements
- Subliminal Steering: Stronger Encoding of Hidden Signals
- MedStruct-S: A Benchmark for Key Discovery, Key-Conditioned QA and Semi-Structured Extraction from OCR Clinical Reports
- The End of Transformers? On Challenging Attention and the Rise of Sub-Quadratic Architectures
- Untangling the Mechanisms of Misleading Context in Medical Question Answering