Inside the Black Box: Cracking AI and Deep Learning

Inside the Black Box: Cracking AI and Deep Learning

Jellypod

Episodes 29
Avg. Duration 12m
Activity Highly Active
Since Nov 2025
Latest Episode May 2026

Publishing Details

Schedule
Weekly
Format
Episodic
Consistency
64%
Hosting
arshavir.jellypod.com

Contact & Outreach

About This Podcast

How do Large Language Models like ChatGPT work, anyway? (Powered by Jellypod)

Podcasting 2.0 Features

episode generator transcript

Explore Statistics

Recent Episodes

Fine Tuning Lora: It's Not What You Think

May 15, 2026 15m Transcript

When you fine-tune an AI model, what changes inside doesn't predict what changes outside. This week on Inside the Black Box, I break down why — and what it means for anyone auditing or regulating…

When Fluent Answers Start Sounding True

May 02, 2026 15m Transcript

This episode explores why smooth, coherent language can feel more credible than it is, and how processing fluency, familiarity, and authority cues shape what we believe. It also digs into why…

Why Your Brain Believes the Model

Apr 27, 2026 24m Transcript

The Heuristic Loop You Can't Break from Inside

When Polished Answers Feel Finished

Apr 20, 2026 27m Transcript

This episode explores fluency-as-validity: the way polished AI responses can make us feel like the work of judgment is already done. It also looks at why large language models are so effective at…

What Seneca Teaches Us that Marcus Couldn't

Apr 12, 2026 17m Transcript

716 features fire on both Seneca and Marcus Aurelius but stay dark for ad copy. The model learned Stoic philosophy, not just an author's style. Plus: why 'inert' features aren't all the same thing.

The Pattern Holds for Another Author

Apr 04, 2026 15m Transcript

We trained a fresh LoRA on the letters of Seneca and ran the same analysis pipeline we used on Marcus Aurelius and advertising copy. Every structural finding replicated. The model organizes its…

The Pattern Holds

Mar 30, 2026 18m Transcript

We replicated our Marcus Aurelius findings at a new layer, then threw the whole method at 12 commercial ad copy styles trained into a single LoRA. The patterns held, and the new domain revealed…

Cracking Open the Black Box

Mar 22, 2026 11m Transcript

We opened the 65%. The features that resisted interpretation one at a time turned out to organize into five co-activation clusters with clear thematic identities and causal effects nearly ten times…

Inside a Fine-Tuned Language Model

Mar 12, 2026 18m Transcript

A concise, single-segment episode of Inside the Black Box: Cracking AI and Deep Learning where Arshavir Blackwell explains, in one continuous narrative, what neural networks are, how their simple…

What Counts as Structure? From Harris and Elman to Today’s Neural Nets

Mar 06, 2026 13m Transcript

This episode of Inside the Black Box: Cracking AI and Deep Learning tells the story of an unexpected convergence in the history of language and AI. In 1995, Peter Bensch noticed that Zelig Harris, a…

Building a House Without Blueprints: When Interpretability Tools Work — and When They Don’t

Feb 27, 2026 18m Transcript

This episode of Inside the Black Box: Cracking AI and Deep Learning explores a new theoretical framework that unifies sparse autoencoders (SAEs), transcoders, and crosscoders — and what it tells us…

I Told My LLM Not to Say "Empower"

Feb 19, 2026 6m Transcript

In this episode of Inside the Black Box: Cracking AI and Deep Learning, Arshavir Blackwell, PhD, takes engineers and researchers inside the practical mechanics of LoRA, low‑rank adaptation methods…

Beyond the Surface of AI Intelligence

Feb 09, 2026 14m Transcript

This episode dives into why judging AI by behavior alone falls short of proving true intelligence. We explore how insights from mechanistic interpretability and cognitive science reveal what’s really…

Unlocking BERTs Hidden Grammar

Feb 03, 2026 9m Transcript

Explore how BERT’s attention heads reveal an emergent understanding of language structure without explicit supervision. Discover the role of attention as a form of memory and what it means for the…

Cracking the Code of AI Interpretation

Jan 28, 2026 10m Transcript

Dive into how we naturally explain neural networks with folk interpretability and why these simple stories fall short. Discover the journey toward mechanistic understandability in AI and what that…

Decoding GPTs Hidden Circuits

Jan 26, 2026 11m Transcript

Explore how sparse autoencoders and transcoders unveil the inner workings of GPT-2 by revealing functional features and computational circuits. Discover breakthrough methods that shift from observing…

Decoding Attention and Emergence in AI

Jan 14, 2026 6m Transcript

Explore how attention heads uncover patterns through learned queries and keys, revealing emergent behaviors shaped by optimization. Dive into parallels with natural selection and psycholinguistics to…

When Knowledge Battles Noise in GPT Models

Jan 07, 2026 7m Transcript

Explore how GPT-2 balances fleeting factual recall with generic responses through internal competition among candidate answers. Discover parallels with human cognition and how larger models navigate…

Inside Circuits: How Large Language Models Understand

Jan 01, 2026 7m Transcript

Dive into the world of neural circuits within large language models. In this episode, Arshavir Blackwell unpacks how transformer circuits, attention mechanisms, and high-dimensional geometry combine…

Hallucinations, Interpretability, and the Seahorse Mirage

Dec 29, 2025 9m Transcript

This episode dives into why advanced language models still generate hallucinations, how interpretability tools help us uncover their hidden workings, and what the seahorse emoji teaches us about…

Frequently Asked Questions

How many episodes does Inside the Black Box: Cracking AI and Deep Learning have?

Inside the Black Box: Cracking AI and Deep Learning has published 29 episodes since November 2025, covering topics in Education, Technology.

Is Inside the Black Box: Cracking AI and Deep Learning still active?

Inside the Black Box: Cracking AI and Deep Learning is currently highly active with new episodes weekly. Average episode length is 12m.

How do I contact Inside the Black Box: Cracking AI and Deep Learning for sponsorship or guest appearances?

Sign up on Grep.FM to access contact details for Inside the Black Box: Cracking AI and Deep Learning, including email and social media links.

Similar Podcasts