# Understanding AI Evaluation Patterns: How Different GPT Models Assess Vision-Language Descriptions

arXiv: https://arxiv.org/abs/2509.10707 · HTML: https://mlsanity.com/paper/2509.10707/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 12 Sep 2025.

## Built on by (most cited first, 1 of 1)

- [Single LLM Debate, MoLaCE: Mixture of Latent Concept Experts Against Confirmation Bias](https://mlsanity.com/paper/2512.23518/) (2025)

## Abstract (arXiv)

As AI systems increasingly evaluate other AI outputs, understanding their assessment behavior becomes crucial for preventing cascading biases. This study analyzes vision-language descriptions generated by NVIDIA's Describe Anything Model and evaluated by three GPT variants (GPT-4o, GPT-4o-mini, GPT-5) to uncover distinct "evaluation personalities" the underlying assessment strategies and biases each model demonstrates. GPT-4o-mini exhibits systematic consistency with minimal variance, GPT-4o excels at error detection, while GPT-5 shows extreme conservatism with high variability. Controlled experiments using Gemini 2.5 Pro as an independent question generator validate that these personalities are inherent model properties rather than artifacts. Cross-family analysis through semantic similarity of generated questions reveals significant divergence: GPT models cluster together with high similarity while Gemini exhibits markedly different evaluation strategies. All GPT models demonstrate a consistent 2:1 bias favoring negative assessment over positive confirmation, though this pattern appears family-specific rather than universal across AI architectures. These findings suggest that evaluation competence does not scale with general capability and that robust AI assessment requires diverse architectural perspectives.