Dual Traits in Probabilistic Reasoning of Large Language Models

Shenxiong Li, Huaxia Rui
12/15/2024
cs.AIcs.CLcs.CY

No official code found. Best implementation: Jacobsonradical/LLM-Dual-Trait (0 stars, last commit 2024-12-15).

Abstract

We conducted three experiments to investigate how large language models (LLMs) evaluate posterior probabilities. Our results reveal the coexistence of two modes in posterior judgment among state-of-the-art models: a normative mode, which adheres to Bayes' rule, and a representative-based mode, which relies on similarity -- paralleling human System 1 and System 2 thinking. Additionally, we observed that LLMs struggle to recall base rate information from their memory, and developing prompt engineering strategies to mitigate representative-based judgment may be challenging. We further conjecture that the dual modes of judgment may be a result of the contrastive loss function employed in reinforcement learning from human feedback. Our findings underscore the potential direction for reducing cognitive biases in LLMs and the necessity for cautious deployment of LLMs in critical areas.

View on arXivView PDF

Code Implementations(1)

Jacobsonradical/LLM-Dual-TraitCommunity100%

This is the official data and code repository for the paper "Dual Traits in Probabilistic Reasoning of Large Language Models"

00PythonOct 3, 20241 years ago

Similar Papers

Incoherent Probability Judgments in Large Language Models

86% match2024

Emergent Bayesian Behaviour and Optimal Cue Combination in LLMs

86% match2025

Large Language Models Reasoning Abilities Under Non-Ideal Conditions After RL-Fine-Tuning

86% match2025

Large Language Models are Biased Reinforcement Learners

86% match1 repo42024

Do Language Models Have Bayesian Brains? Distinguishing Stochastic and Deterministic Decision Patterns within Large Language Models

85% match2025

Forgetting as a Feature: Cognitive Alignment of Large Language Models

85% match2025

Cite this paper

@article{li2024dual,
  title  = {Dual Traits in Probabilistic Reasoning of Large Language Models},
  author = {Shenxiong Li and Huaxia Rui},
  year   = {2024},
  eprint = {2412.11009},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2412.11009v1}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2412.11009.svg)](https://codeofpaper.com/paper/2412.11009)

Discussion