Exploring the Biology of LLMs with Circuit Tracing with Emmanuel Ameisen - #727

Exploring the Biology of LLMs with Circuit Tracing with Emmanuel Ameisen - #727

In this episode, Emmanuel Ameisen, a research engineer at Anthropic, returns to discuss two recent papers: "Circuit Tracing: Revealing Language Model Computational Graphs" and "On the Biology of a Large Language Model." Emmanuel explains how his team developed mechanistic interpretability methods to understand the internal workings of Claude by replacing dense neural network components with sparse, interpretable alternatives. The conversation explores several fascinating discoveries about large language models, including how they plan ahead when writing poetry (selecting the rhyming word "rabbit" before crafting the sentence leading to it), perform mathematical calculations using unique algorithms, and process concepts across multiple languages using shared neural representations. Emmanuel details how the team can intervene in model behavior by manipulating specific neural pathways, revealing how concepts are distributed throughout the network's MLPs and attention mechanisms. The discussion highlights both capabilities and limitations of LLMs, showing how hallucinations occur through separate recognition and recall circuits, and demonstrates why chain-of-thought explanations aren't always faithful representations of the model's actual reasoning. This research ultimately supports Anthropic's safety strategy by providing a deeper understanding of how these AI systems actually work. The complete show notes for this episode can be found at https://twimlai.com/go/727.

Episoder(781)

Is It Time to Rethink LLM Pre-Training? with Aditi Raghunathan - #747

Is It Time to Rethink LLM Pre-Training? with Aditi Raghunathan - #747

Today, we're joined by Aditi Raghunathan, assistant professor at Carnegie Mellon University, to discuss the limitations of LLMs and how we can build more adaptable and creative models. We dig into her...

16 Sep 202558min

Building an Immune System for AI Generated Software with Animesh Koratana - #746

Building an Immune System for AI Generated Software with Animesh Koratana - #746

Today, we're joined by Animesh Koratana, founder and CEO of PlayerZero to discuss his team’s approach to making agentic and AI-assisted coding tools production-ready at scale. Animesh explains how rap...

9 Sep 20251h 5min

Autoformalization and Verifiable Superintelligence with Christian Szegedy - #745

Autoformalization and Verifiable Superintelligence with Christian Szegedy - #745

In this episode, Christian Szegedy, Chief Scientist at Morph Labs, joins us to discuss how the application of formal mathematics and reasoning enables the creation of more robust and safer AI systems....

2 Sep 20251h 11min

Multimodal AI Models on Apple Silicon with MLX with Prince Canuma - #744

Multimodal AI Models on Apple Silicon with MLX with Prince Canuma - #744

Today, we're joined by Prince Canuma, an ML engineer and open-source developer focused on optimizing AI inference on Apple Silicon devices. Prince shares his journey to becoming one of the most prolif...

26 Aug 20251h 10min

Genie 3: A New Frontier for World Models with Jack Parker-Holder and Shlomi Fruchter - #743

Genie 3: A New Frontier for World Models with Jack Parker-Holder and Shlomi Fruchter - #743

Today, we're joined by Jack Parker-Holder and Shlomi Fruchter, researchers at Google DeepMind, to discuss the recent release of Genie 3, a model capable of generating “playable” virtual worlds. We dig...

19 Aug 20251h 1min

Closing the Loop Between AI Training and Inference with Lin Qiao - #742

Closing the Loop Between AI Training and Inference with Lin Qiao - #742

In this episode, we're joined by Lin Qiao, CEO and co-founder of Fireworks AI. Drawing on key lessons from her time building PyTorch, Lin shares her perspective on the modern generative AI development...

12 Aug 20251h 1min

Context Engineering for Productive AI Agents with Filip Kozera - #741

Context Engineering for Productive AI Agents with Filip Kozera - #741

In this episode, Filip Kozera, founder and CEO of Wordware, explains his approach to building agentic workflows where natural language serves as the new programming interface. Filip breaks down the ar...

29 Jul 202546min

Infrastructure Scaling and Compound AI Systems with Jared Quincy Davis - #740

Infrastructure Scaling and Compound AI Systems with Jared Quincy Davis - #740

In this episode, Jared Quincy Davis, founder and CEO at Foundry, introduces the concept of "compound AI systems," which allows users to create powerful, efficient applications by composing multiple, o...

22 Jul 20251h 13min

Populært innen Politikk og nyheter

giver-og-gjengen-vg
aftenpodden
aftenpodden-usa
stopp-verden
forklart
i-retten
popradet
lydartikler-fra-aftenposten
fotballpodden-2
rss-gukild-johaug
det-store-bildet
dine-penger-pengeradet
rss-ness
nokon-ma-ga
hanna-de-heldige
aftenbla-bla
frokostshowet-pa-p5
rss-dannet-uten-piano
grasoner-den-nye-kalde-krigen
e24-podden