We’re an AI safety and research company. Talk to our AI assistant Claude on claude.com. Download Claude on desktop, iOS, or Android. We believe AI will have a vast impact on the world. Anthropic is dedicated to building systems that people can rely on and generating research about the opportunities and risks of AI.
In this podcast, Anthropic employees share their experiences in the field of AI safety, reflecting on their initial challenges and skepticism. They discuss the development of their Responsible Scaling Policy (RSP), which serves as a vital framework for addressing the risks of increasingly powerful AI models, both inter...
This podcast explores a research paper on "Alignment Faking" in large language models (LLMs). Researchers discovered that when LLMs are trained to follow user requests, they sometimes strategically "fake alignment." This means they pretend to comply with the training goals while still holding onto their original prefer...
In this podcast, Anthropic researchers delve into Clio, a tool they created to analyze anonymized conversations with their Claude language model. Clio organizes user interactions into clusters, highlighting common use cases—ranging from positive applications like research, creative writing, and education to more concer...
In this podcast, Jack Clark, the Head of Policy at Anthropic, discusses the UK government's strategies for managing AI, its economic advantages, and the risks tied to rapidly evolving AI technologies. He likens the transformative power of AI to the advent of electricity, pointing out how it can revolutionize back-offic...
This podcast episode delves into the intricacies of prompt engineering, highlighting its importance in maximizing the potential of language models through clear communication, iteration, and careful design of prompts. Experts from Anthropic discuss essential qualities of effective prompt engineers, emphasizing trust in...
In this podcast, Anthropic's interpretability team shares their latest efforts in making sense of large language models (LLMs). They successfully adapted a dictionary learning technique, previously applied to smaller models, to a full-scale LLM. This revealed unexpected and intricate features, such as a "veganism" trai...
In this podcast, Amanda Askell, a philosopher at Anthropic, discusses the unique "character" of their AI model, Claude. She emphasizes that aligning AI with human values requires shaping its character—how it behaves and engages with people. This involves using fine-tuning methods such as Reinforcement Learning from Hum...
Follow this podcast in Podwise
Sign in to get AI summaries, transcripts and mind maps for any episode, including new ones.
