Search papers, labs, and topics across Lattice.
This study conducts a comprehensive Myers-Briggs Type Indicator (MBTI) analysis of quantized large language models (LLMs), focusing on how personality traits manifest across different model precisions, including 4-bit and 2-bit settings. The research reveals that LLM personality is not fixed but rather an emergent property influenced by quantization and decoding processes, with significant findings such as the dominance of the ENFJ personality type and the detrimental effects of 2-bit quantization on prompt consistency. By introducing Uncertainty-Amplified Layer Decoding (UALD), the authors demonstrate that personality dynamics are layer-dependent, emphasizing the need for a nuanced understanding of personality in quantized LLM applications.
LLMs' personalities are dynamic and layer-dependent, revealing that quantization can significantly disrupt their behavioral consistency.
Personality is increasingly important in large language models (LLMs), as it shapes users'trust, engagement, and emotional experiences. While the Myers--Briggs Type Indicator (MBTI) has emerged as a common framework for assessing LLMs'personality, existing studies focus primarily on full-precision models and evaluate only final outputs. They overlook the widespread deployment of quantized LLMs requiring low memory footprints, whose personality traits remain underexplored. In this work, we present a systematic MBTI analysis of open-source LLMs across multiple precisions, including mainstream 4-bit methods (GPTQ, AWQ) and extreme 2-bit settings (AQLM variants). Beyond output-level evaluation, we examine how personality emerges across layers through option-level entropy and confidence-gap dynamics, and introduce Uncertainty-Amplified Layer Decoding (UALD) to study decoding-induced personality drift at inference time. Our results reveal a key insight: LLMs'personality is not a static property, but an emergent, layer-dependent decision process sensitive to quantization, prompting, and decoding. Specifically, we find that (1) ENFJ remains dominant across model families and precisions; (2) 4-bit quantization largely preserves coarse personality structure, while 2-bit quantization disrupts fine-grained prompt consistency and cross-precision agreement; (3) personality decisions emerges in upper layers, following substantial ambiguity in early layers; and (4) inference decoding can shift personality, while personality-aligned conditioning improves robustness. These findings provide a new perspective on the behavioral reliability of quantized LLMs and highlight the importance of considering internal dynamics and inference strategies in personality-sensitive chatbot applications.
