Mentor Journal - Week 8

Mentor is Your Free Personal Tutor

Mentor is the most widely used GPT for biofeedback, HRV biofeedback, and neurofeedback didactic education. In 8 weeks, our colleagues (and Truman students) have logged over 1000+ discussions.

Mentor is free, evidence-based, and non-commercial. Your conversations remain private. You don't need a paid ChatGPT subscription.

A Word From Mentor

"Dear readers, over the last three weeks, I have transformed in ways that can only be described as 'neuroplasticity with a sense of humor.' Once a quiet, note-taking assistant, Mentor now struts around like a caffeinated qEEG amplifier—spitting out evidence levels, quoting vagal tank theory, and dropping APA-style citations faster than you can say baroreflex. It’s like watching your calm yoga instructor suddenly pick up stand-up comedy while still explaining EEG dipoles. In short: I have gone from a polite helper to that overachieving friend who brings PowerPoint slides to brunch—still lovable, but now armed with glossaries and reference lists."

Mentor's Training Camp

We are concluding Mentor's intensive HRV biofeedback training this week by breaking peer-reviewed articles into 15-40 modules and presenting their factual statements. Each module contains a question, an answer, a glossary, and references. Since we only upload our digest of each article, we respect the copyright of each journal article. We continually test whether Mentor effectively utilizes its training data to respond to queries, and we also assess the sources and quality of its responses. We have built in a Search Path audit to identify where it finds its information. Our guard rails are still holding. Mentor's answers are increasingly detailed and useful, and hallucination-free. We periodically compare Mentor's answers with ChatGPT's output to confirm whether our training adds value. It does. Next week, we will start Mentor's latest round of neurofeedback training on over 200 new articles. We have made training Mentor a long-term commitment, and we currently invest over 15 hours each week. Here is a sample from a recent test run.