Eric Bigelow · "The Cognitive Revolution" · How Agents Decide: Goodfire's Eric Bigelow on Critical Tokens, Phase Shifts, & In-Context Learning
Audio streams from the publisher’s own feed. Verbatim does not host or copy it.
Use high-sampling of model outputs to spot smooth curves, then switch to a statistical model for efficiency
“What we tried to do is, you know, using large language models and AI, we tried to really ground probability of success in an empiric data set, where we use Bayesian statistics.”
“I would do cost by app… or cost by model… I track spend and efficiency by model, so devs can look and see what’s the best model per their workload”
“We have a bunch of fine-tuned models ourselves too… making us more innovative… couldn’t do with the frontier models because of cost and efficiency reasons”
“The more you post-train models the less diverse their outputs become”
The strongest new moments on your topics, in the speakers' own words. Free for three topics.
No card. No password. One-click unsubscribe.
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.