Explore how CTGT’s research is solving the problems standard RAG and fine-tuning cannot touch.

Dec 9, 2025

Introducing Mentat: Deterministic Runtime Control for LLMs

Today, we’re excited to introduce Mentat, CTGT’s new OpenAI-compatible endpoint built to deliver something the AI industry has been missing: deterministic control over large language models.

More Blogs

Research, publications & industry insights

Mar 14, 2025

Research

A Feature-Level Approach to Mitigating Bias and Censorship in DeepSeek-R1

Feb 20, 2025

Announcement

CTGT Raises $7.2M To Help Enterprises Break Through The Limits Of AI Compute

Jan 8, 2024

Research

Empirical Analysis of Efficient Fine-Tuning Methods for Large Pre-Trained Language Models

May 6, 2023

Research

Explore how CTGT’s research is solving the problems standard RAG and fine-tuning cannot touch.

Introducing Mentat: Deterministic Runtime Control for LLMs

Research, publications & industry insights

A Feature-Level Approach to Mitigating Bias and Censorship in DeepSeek-R1

CTGT Raises $7.2M To Help Enterprises Break Through The Limits Of AI Compute

Empirical Analysis of Efficient Fine-Tuning Methods for Large Pre-Trained Language Models

Training data eigenvector dynamics in the EigenPro implementation of the neural tangent kernel and recursive feature machines.

Experience the next  evolution of AI Governance

Explore how CTGT’s research is solving the problems standard RAG and fine-tuning cannot touch.

Introducing Mentat: Deterministic Runtime Control for LLMs

Research, publications & industry insights

A Feature-Level Approach to Mitigating Bias and Censorship in DeepSeek-R1

CTGT Raises $7.2M To Help Enterprises Break Through The Limits Of AI Compute

Empirical Analysis of Efficient Fine-Tuning Methods for Large Pre-Trained Language Models

Training data eigenvector dynamics in the EigenPro implementation of the neural tangent kernel and recursive feature machines.

Experience the next evolution of AI Governance

Experience the next  evolution of AI Governance