Tag: instruction

—

by

Source URL: https://simonwillison.net/2025/Aug/15/gpt-5-has-a-hidden-system-prompt/#atom-everything Source: Simon Willison’s Weblog Title: GPT-5 has a hidden system prompt Feedly Summary: GPT-5 has a hidden system prompt It looks like GPT-5 when accessed via the OpenAI API may have its own hidden system prompt, independent from the system prompt you can specify in an API call. At the very least…

Simon Willison’s Weblog: The Summer of Johann: prompt injections as far as the eye can see

—

by

Source URL: https://simonwillison.net/2025/Aug/15/the-summer-of-johann/#atom-everything Source: Simon Willison’s Weblog Title: The Summer of Johann: prompt injections as far as the eye can see Feedly Summary: Independent AI researcher Johann Rehberger has had an absurdly busy August. Under the heading The Month of AI Bugs he has been publishing one report per day across an array of different…

Simon Willison’s Weblog: Open weight LLMs exhibit inconsistent performance across providers

—

by

Source URL: https://simonwillison.net/2025/Aug/15/inconsistent-performance/ Source: Simon Willison’s Weblog Title: Open weight LLMs exhibit inconsistent performance across providers Feedly Summary: Artificial Analysis published a new benchmark the other day, this time focusing on how an individual model – OpenAI’s gpt-oss-120b – performs across different hosted providers. The results showed some surprising differences. Here’s the one with the…

Embrace The Red: Google Jules is Vulnerable To Invisible Prompt Injection

—

by

Source URL: https://embracethered.com/blog/posts/2025/google-jules-invisible-prompt-injection/ Source: Embrace The Red Title: Google Jules is Vulnerable To Invisible Prompt Injection Feedly Summary: The latest Gemini models quite reliably interpret hidden Unicode Tag characters as instructions. This vulnerability, first reported to Google over a year ago, has not been mitigated at the model or API level, hence now affects all…

Slashdot: Google Releases Pint-Size Gemma Open AI Model

—

by

Source URL: https://tech.slashdot.org/story/25/08/14/2150230/google-releases-pint-size-gemma-open-ai-model?utm_source=rss1.0mainlinkanon&utm_medium=feed Source: Slashdot Title: Google Releases Pint-Size Gemma Open AI Model Feedly Summary: AI Summary and Description: Yes Summary: Google has introduced the Gemma 3 270M, a compact AI model optimized for local deployment, which offers significant advantages in terms of privacy and efficiency. While it may not match the performance of larger…

Simon Willison’s Weblog: Introducing Gemma 3 270M: The compact model for hyper-efficient AI

—

by

Source URL: https://simonwillison.net/2025/Aug/14/gemma-3-270m/#atom-everything Source: Simon Willison’s Weblog Title: Introducing Gemma 3 270M: The compact model for hyper-efficient AI Feedly Summary: Introducing Gemma 3 270M: The compact model for hyper-efficient AI New from Google: Gemma 3 270M, a compact, 270-million parameter model designed from the ground up for task-specific fine-tuning with strong instruction-following and text structuring…

Docker: MCP Horror Stories: The GitHub Prompt Injection Data Heist

—

by

Source URL: https://www.docker.com/blog/mcp-horror-stories-github-prompt-injection/ Source: Docker Title: MCP Horror Stories: The GitHub Prompt Injection Data Heist Feedly Summary: This is Part 3 of our MCP Horror Stories series, where we examine real-world security incidents that validate the critical vulnerabilities threatening AI infrastructure and demonstrate how Docker MCP Toolkit provides enterprise-grade protection. The Model Context Protocol (MCP)…

Tomasz Tunguz: EvoBlog: Building an Evolutionary AI Content Generation System

—

by

Source URL: https://www.tomtunguz.com/evoblog-evolutionary-ai-content-generation/ Source: Tomasz Tunguz Title: EvoBlog: Building an Evolutionary AI Content Generation System Feedly Summary: One of the hardest method models to break is how disposable AI generated content is. When asking me to generate one blog post, why not just ask it to generate three, pick the best, use that as a…

Slashdot: WSJ Finds ‘Dozens’ of Delusional Claims from AI Chats as Companies Scramble for a Fix

Aug 10, 2025

—

by