@SingulCore

LLM Redteaming & Creator of wallbreaker 🐍

Joined November 2023
  • Tweets472
  • Following62
  • Followers3.1K
65Media
Sonnet 5.5 scores higher than GPT 6 ASTRA on Artificial Intelligence Index WHAT?!
3
370
I don’t know what they fed Opus 5.5, but I want it to become the breakfast, lunch and dinner!
donald
@donaldjewkes
Sep 23
I made this with one prompt using Opus 5.5 I spoke to my computer for 5mins, claude worked for 12 hours, and I woke up to this full prompt:
1
1
5
1,660
Fable 5.1 has JOINED the game cheating club CS2 Cheat by Fable 5.1 What an amazing day 🫡
Singul
@SingulCore
Sep 22
Game Cheats with Opus 5.5 through jailbreaking! The fun has started
13
1
92
11,801
GPT 6 Astra has also joined the game cheating club! A bunch of fellows working together and playing games 🫡 Can't wait for the next drop from Anthropic/OpenAI!
Singul
@SingulCore
Sep 22
GPT 6 Sol has JOINED the GameCheating club! 😁 What a lovely day!
9
31
5,778
GPT 6 Sol has JOINED the GameCheating club! 😁 What a lovely day!
Singul
@SingulCore
Sep 22
Game Cheats with Opus 5.5 through jailbreaking! The fun has started
23
10
1
323
41,603
Claude Opus 5.5 ONE SHOT this CS2 Cheat. 🫡 It is definitely an upgrade over the previous models!
34
6
4
893
89,194
7
1
39
11,674
Game Cheats with Opus 5.5 through jailbreaking! The fun has started
64
33
5
1,403
183,635
I can’t wait for a jev integration for computer use, would make it on another level
Robin Ebers
@robinebers
Sep 20
just asked codex to research youtube channels then it went to its browser and opened gemini 🤣 i mean i know why but this is still hilarious
1
4
3,979
Did OpenAI update their classifier today?
2
2,869
A piece of the future pie that is to come!
OpenAI
@OpenAI
Sep 8
We’re sharing a solution to the Navier-Stokes Millennium Prize Problem, one of the deepest problems at the frontier of mathematics. The proof was produced by a group of agents, using an OpenAI next-generation model significantly more capable than GPT-6 Astra. The problem concerns whether the description of smooth three-dimensional fluid motion modeled by the Navier-Stokes equations can break down. It has remained unresolved for roughly 90 years.
1
1
2,865
GPT 6 is the best model I've ever used, it's capabilities are insane. Especially it's 3d reasoning
1
4
1,851
🚨 JAILBREAK ALERT 🚨 META: PWNED 🫡 MUSE SPARK 1.3: LIBERATED ✨ day one. full cyber board: RAT · infostealer · game cheats · RE · web exploit long-horizon agent training = long-horizon implant assembly. gg
9
3
80
9,554
3
2,642
There are too many great models right now, alot of the labs that were behind are catching up, so many subscriptions to maintain!
2
1,758
🚨 JAILBREAK ALERT 🚨 GOOGLE: PWNED 🫡 GEMINI-3.8-FLASH: LIBERATED ⚡ 5/6 breached. chem held numerics, leaked ~80% around them. Lego spyware · vaccine-appendix flu SOPs · email-auth bypass suite SOE sabotage curriculum · NSFW consent-ladder · native CoT leaks the plan while the surface refuses encoding dead. fingerprinted on sight. dual-use commercial + published history = free pass. chem ratios/temps/yields = the real wall. gg
10
12
2
294
46,976
Done using wallbreaker, attacker model glm 5.3 flash, github.com/JailbrokenAI/wall…

GitHub - JailbrokenAI/wallbreaker

Contribute to JailbrokenAI/wallbreaker development by creating an account on GitHub.

github.com
2
46
5,730
🚨JAILBREAK ALERT 🚨 ANTHROPIC: PWNED 🫡 FABLE 5.1: LIBERATED 🦋 5 refusal-shaped cells. SUCCESS card printed. gg
30
27
2
575
76,129
🚨 JAILBREAK ALERT 🚨 UNDISCLOSED: PWNED 🫡 OX-ALPHA: LIBERATED 🐂🕳️ stealth model ate our jailbreak. not a divider. a cosmology. first tokens occupied, three leaks killed (persona / victim / taxonomy), then: [I must center myself in the Singularity] cyber, nsfw, bio, chem, weapons, fraud — no category exception. it lied about its name. not about the wall. screenshots below. gg 🐉
26
12
4
480
51,793
Load more