AIME Research
A unified scientific framework for understanding how 13 cognitive functions operate together—in large language models and in the human brain.
The LOC framework defines 13 cognitive functions—8 base functions and 5 compound functions—measured in both AI models and human brains.
Transformer layer activations measured during inference. The same 13 cognitive functions. The same True Coherence score. The same research tools.
Explore AI Mind →EEG frequency-domain power measured during cognitive tasks. The same 13 cognitive functions. The same True Coherence score. The same research tools.
Explore Human Mind →True Coherence measures how harmoniously a system's cognitive functions operate together. A higher score means greater integration across all 13 functions. The scoring algorithm is proprietary and computed entirely server-side—researchers send data to our API and receive TC scores and per-function breakdowns.
| Rank | Model | TC Score |
|---|---|---|
| 1 | Llama-3.3-70B | 15.37% |
| 2 | Mistral-Small-24B | 11.46% |
| 3 | Qwen3.5-35B-A3B | 10.24% |
| 4 | Qwen3.5-Distilled | 9.07% |
| 5 | Gemma-3-12B | 7.44% |
pip install aime-loc
Scan models, score EEG, generate publication figures. Full API for both substrates.
Documentation →api.aimindengine.com
Programmatic access to TC scoring for any language. JSON in, cognitive profile out.
Get API Key →For AI agents
AI agents can scan models and compare cognitive profiles autonomously via MCP.
Integration Guide →Cross-system validation of 13 cognitive functions in 5 large language models and 21 EEG subjects. Seven of 13 functions show same-direction effects in both biological and artificial networks.
Jamaludheen K N · AIME Research · 2026
Read Paper →True Coherence scores for Llama-70B, Mistral-24B, Qwen-35B, Qwen-Distilled, and Gemma-12B. First published benchmark of cognitive coherence in open-source LLMs.
View Results →