Loading profile…

World & Business

Home
Technology
Gaming
Entertainment
World & Business
Science
Sports
AI

Home
Technology
Gaming
Entertainment
World & Business
Science
Sports
AI

David Reber (@davidpreber) · Digg

David Reber

@davidpreber

#AIsafety, interpretability, and causality. PhD student in CS @ #UChicago.

Featured in

AI

New paper examines which parts of language-model rationales actually matter

A post uses an API-price forecast to argue that only contrasting forecasts can validate which points matter to its conclusion.

3 Sources, 3h ago, first seen 3h ago

Flowchart. One forecasting question, whether top-model API prices halve by 2027, fans out into 30 forecasts from Claude Opus 5.5, ranging 7-15% yes. Pairs of forecasts are contrasted, each pair yielding crux questions. Paraphrases are merged into a column of 86 cruxes. Headline: cruxes are properties of the distribution of forecasts, not of any one forecast.

Connections

@davidpreber