AIvengo > Reviews > Anthropic accidentally revealed an internal document about Claude’s “soul”

Anthropic accidentally revealed an internal document about Claude’s “soul”

Anthropic accidentally revealed the “soul” of artificial intelligence to a user. And this is not a metaphor. This is a quite specific internal document.

User Richard Weiss inadvertently made the large language model Claude 4 and 5 Opus quote a document called “Soul Overview”. Anthropic’s philosopher and ethics specialist Amanda Askell confirmed its authenticity. It was uploaded during the training phase.

Weiss requested from Claude the system message with instructions for conducting dialogue. The chatbot referenced several documents, 1 of which is called “soul overview”. The user asked to show the text, and Claude produced an 11 thousand word guide on how the model should behave.

The document contains safety instructions with protective barriers against dangerous responses. Claude is tasked with being genuinely helpful to people.

Anthropic’s ethics specialist Amanda Askell confirmed, I quote: “I’ve been working on it for some time, it’s still in the refinement stage. And soon we plan to release a full version with more detailed information. Model quotes aren’t particularly accurate, but the original document is mostly correct. In the company it’s called the soul document, and Claude obviously too. But we’ll call it something else”.

It turns out, Claude’s “Soul” turned out to be a behavior manual, and an ordinary user gained access. Many details of AI model development remain behind the scenes. And the opportunity to peek behind them — a small surprise.

Autor: AIvengo

For 5 years I have been working with machine learning and artificial intelligence. And this field never ceases to amaze, inspire and interest me.

UBTech will send Walker S2 robots to serve on China's border for $37 million

Chinese company UBTech won a contract for $37 million. And will send humanoid robots Walker S2 to serve on China's border with Vietnam. South China Morning Post reports that the robots will interact with tourists and staff, perform logistics operations, inspect cargo and patrol the area. And characteristically — they can independently change their battery.

Anthropic accidentally revealed an internal document about Claude's "soul"

Anthropic accidentally revealed the "soul" of artificial intelligence to a user. And this is not a metaphor. This is a quite specific internal document.

Jensen Huang ordered Nvidia employees to use AI everywhere

Jensen Huang announced total mobilization under the banner of artificial intelligence inside Nvidia. And this is no longer a recommendation. This is a requirement.

AI chatbots generate content that exacerbates eating disorders

A joint study by Stanford University and the Center for Democracy and Technology showed a disturbing picture. Chatbots with artificial intelligence pose a serious risk to people with eating disorders. Scientists warn that neural networks hand out harmful advice about diets. They suggest ways to hide the disorder and generate "inspiring weight loss content" that worsens the problem.

OpenAGI released the Lux model that overtakes Google and OpenAI

Startup OpenAGI released the Lux model for computer control and claims this is a breakthrough. According to benchmarks, the model overtakes analogues from Google, OpenAI and Anthropic by a whole generation. Moreover, it works faster. About 1 second per step instead of 3 seconds for competitors. And 10 times cheaper in cost per processing 1 token.