“In short, Kurt shows that you can ablate roughly 90% of the individual neurons, keeping only those with the strongest activations, with surprisingly minimal performance degradation.”
“The authors do show that a very small amount of additional post-training on the sparsely activating model is enough to recover a large majority of the performance loss.”
Kurt Tigges · "The Cognitive Revolution" · The Neuron Basis: Goodfire's Curt Tigges on Agentic Interp, Monosemanticity & Reverse Engineering
Audio streams from the publisher’s own feed. Verbatim does not host or copy it.
What was described
performed a small amount of additional post-training on the sparsely activating model to recover performance loss after heavy ablation
Said nearby on the same subject
“Showing that the original model's behavior is substantially restored relative to ablating the neuron entirely when the other LLM's guesses are patched in to the model's activations.”
“We're still up an additional 2% year over year. So we're actually tracking growth, not a big amount of growth, but a small amount of growth in the trans-border segment.”
“You are somewhat well known for tweaking the script after you watch the actors perform it”
Moments like this, every Monday
The strongest new moments on your topics, in the speakers' own words. Your first brief is free, then USD {{ TIERS.starter.price }} a month for three topics.
No card to start. No password. Cancel any time.
Track a competitor?
Pro watches up to 25 names or phrases and emails you the morning after they are said.
See Pro, $29 a monthVerbatim quotes a short extract and links to the episode; the episode belongs to its publisher. Remove a show.