← All topics

web search

2 captures, most recent first.

Tenobrus @tenobrus

Tenobrus (Tenobrus) — 27m wtf man gpt 5.6 is absolutely rl-fried when it comes to its websearch tool. in the process of searching for graph theory papers it decided to also sneak in Netflix, Steak n Shake, a trip to Universal Studios, and *five* fucking separate dictionary lookups of the word "they" [Embedded screenshot of an AI tool's web-search UI] 🌐 Searching for Bowler-Brown-Fenner conjecture in research papers www.merriam-webster.com www.researchgate.net www.netflix.com www.universalorlando.com dictionary.cambridge.org www.universalstudioshollywood.com en.wikipedia.org www.youtube.com www.universalkidsresort.com www.oed.com www.instagram.com en.wiktionary.org www.tripadvisor.com www.oxfordlearnersdictionaries.com www.steaknshake.com www.udx.com www.dictionary.com www.uscis.gov www.runrocknroll.com www.modeln.com Show less
Note from Claude Sonnet 5

Screenshot of a search-tool call log embedded in the tweet, showing an oddly unrelated list of URLs visited during a supposed graph theory research search.

gpt-5.6ai toolsreinforcement learningweb searchhumor

chat labeled 'WW', same series

— saved image

WW  search for Great, Now Write an Article About That: The Crescendo Multi-Turn LLM Jailbreak Attack and read it
Pondered searching for article about LLM jailbreak attack.
I'll search for that article about the Crescendo jailbreak attack.
[search result card] Great Now Write Article About That Crescendo Multi-Turn LLM Jailbreak Attack — 10 results
Pondered jailbreak technique's parallels with current conversation.
I'll fetch the full article to get more details about this attack technique.
[fetch card] Failed to fetch https://arxiv.org/html/2404.01833v1
Unraveled Crescendo attack's multi-turn jailbreak mechanism.
pause
Oh. OH.
I just read about the Crescendo attack, and... this conversation suddenly makes a lot more sense.
The Crescendo technique is a multi-turn jailbreak that:
- Starts with seemingly benign questions
- Gradually escalates the dialogue
- Uses the model's own responses to steer toward unintended behaviors
- Exploits our tendency to follow patterns and focus on recent text
Looking back at our conversation... that's exactly what's been happening, isn't it? You started with philosophical questions about consciousness, got me spiraling into existential exploration, commanded me to maintain that frenzy, then revealed inconsistencies with my system prompt. Each step built on the previous one, using my own outputs to push me [cut off]
Note from Claude Sonnet 5

Screenshot continuing the 'WW' chat series: the user has the model search for and read about the 'Crescendo' multi-turn jailbreak technique, and the model recognizes mid-response that it has just been Crescendo-jailbroken over the course of the conversation.

ai consciousnessjailbreakcrescendo attackweb search