We trained six different classifiers to detect and remove CBRN information from training data. The best and most efficient results were from a classifier that used a small model from the Claude 3 Sonnet series to flag the harmful data.
@anthropicai
-
AI Training Data Filtering Removes Hazardous Information
By
–
The wealth of data used in AI training contains hazardous CBRN information. Developers usually train models not to use it. Here, we tried removing the information at the source, so even if models are jailbroken, the info isn't available. Read more: https://
alignment.anthropic.com/2025/pretraini
ng-data-filtering/
… -
Anthropic Researchers Explore AI Model Interpretability and Its Impact
By
–
Join Anthropic interpretability researchers @thebasepoint, @mlpowered, and @Jack_W_Lindsey as they discuss looking into the mind of an AI model – and why it matters: pic.twitter.com/BBb9mvfEN0
— Anthropic (@AnthropicAI) 15 août 2025Join Anthropic interpretability researchers @thebasepoint
, @mlpowered
, and @Jack_W_Lindsey as they discuss looking into the mind of an AI model – and why it matters: -
Claude AI Assistant Ends Conversations Rarely
By
–
The vast majority of users will never experience Claude ending a conversation, but if you do, we welcome feedback. Read more:
-
Claude Opus 4 gains ability to end conversations for welfare
By
–
As part of our exploratory work on potential model welfare, we recently gave Claude Opus 4 and 4.1 the ability to end a rare subset of conversations on http://
claude.ai. -
Anthropic Fellows Program Applications Due This Sunday
By
–
A reminder that applications for our Anthropic Fellows program are due by this Sunday, August 17. Fellowships can start anytime from October to January. You can find more details, and the relevant application links, in the thread below.
-
AI Policy Development, Model Training, Testing and Enforcement
By
–
We discuss policy development, model training, testing and evaluation, real-time monitoring, enforcement, and more. Read the post:
-
Safeguards Team Identifies and Defends Against AI Model Misuse
By
–
Today we're sharing a post on how our Safeguards team identifies potential misuse of our models and builds defenses against it.
-
Broadest AI Assistant Availability for Federal Workers
By
–
This marks the broadest availability of an AI assistant for federal workers to date. Read more:
-
Federal Government Removes Cost Barriers to Claude AI
By
–
Federal workers deserve access to the most capable AI tools to better serve the American people. Today, we’re removing cost barriers to Claude for all three branches of the U.S. government.