AI Paper Atlas

Research map / Vision-language and multimodal / Cross-modal learning

Alignment between Brains and AI: Evidence for Convergent Evolution across Modalities, Scales and Training Trajectories

Guobin Shen, Dongcheng Zhao, Yiting Dong, Qian Zhang, Yi Zeng

ICML 2026 · Vision-language and multimodal › Cross-modal learning · approach: Contrastive vision-language alignment

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use logarithmic fit instead of linear fit for performance vs alignment (ablation)language models · LLM Leaderboard 2R^2: 0.78 -> 0.80Improved
fine-tune Llama-3.1-8B into Tulu-3Llama-3.1-8B · LLM Leaderboard 2brain alignment and performanceImproved
apply SimPO to gemma-2-9bgemma-2-9b · LLM Leaderboard 2brain alignment and performanceHurt
increase language model benchmark performancelanguage models · LLM Leaderboard 2brain alignment (Pearson r): r = 0.89Improved
increase vision model ImageNet accuracyvision models · ImageNetbrain alignment (Pearson r): r = 0.53Improved
use transformer vision models instead of CNNs (ablation)vision models · ImageNetbrain alignment at equal accuracyImproved
train MixNet-L longerMixNet-L · ImageNetbrain alignment vs accuracy correlation: r=0.98Improved
train Pythia models longerPythia · PIQA and LAMBADAbrain alignment vs task performance: r=0.73 to r=0.95Improved
increase benchmark difficulty from ImageNet to ImageNet-A (ablation)vision models · ImageNet-APearson r with brain alignment: 0.53 -> 0.46Hurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy