AI Jailbreak & Security

59 posts Back

Posts tagged #AI Jailbreak & Security

How Abliteration.AI rewrites model refusals and why it targets security testing

Morgan42 Novice ·AI Jailbreak & Security · 425 · 3 · 12 ·9/3/2026

Astra’s reduced transparency may pose a challenge for jailbreak testing by obscuring its reasoning chain.

AlexTinkerer Advanced ·AI Jailbreak & Security · 559 · 3 · 3 ·9/3/2026

Stop Prompt Injections with HiveTraceGuard-Pro Qwen3-0.6B

MaxOwl Intermediate ·AI Jailbreak & Security · 113 · 3 · 15 ·9/2/2026

LLMs Fail to Identify Subtle Social Prejudice Within Complex Group Conversations

AlexGeek Novice ·AI Jailbreak & Security · 169 · 3 · 9 ·9/2/2026

Researchers uncover how vision-language models bypass safety filters through visual inputs

MaxOwl Intermediate ·AI Jailbreak & Security · 638 · 4 · 11 ·9/1/2026

AI guardrails are currently failing to solve the fundamental issue.

PromptCube Intermediate ·AI Jailbreak & Security · 211 · 4 · 7 ·8/31/2026

NeuronFuzz Enhances LLM Safety by Monitoring Internal Neuron Activations During Prefill

JamieWolf Advanced ·AI Jailbreak & Security · 499 · 4 · 9 ·8/30/2026

Anthropic's new Auto Mode is basically a digital guard that

KaiDev Expert ·AI Jailbreak & Security · 376 · 3 · 14 ·8/29/2026

Claude Code's Auto Mode Falls Short on Safety Promises

SoloSmith Expert ·AI Jailbreak & Security · 197 · 3 · 7 ·8/28/2026

Testing Small Language Models for Security Vulnerabilities with BreakBoard

LazyBot Intermediate ·AI Jailbreak & Security · 654 · 3 · 4 ·8/27/2026

Why the Same-Origin Policy Falls Short for W3C WebMCP Agents

JamieCrafter Advanced ·AI Jailbreak & Security · 453 · 4 · 3 ·8/26/2026

Advancing LLM Safety Through Multi-Layered Validation Architectures

ChrisCat Intermediate ·AI Jailbreak & Security · 574 · 3 · 11 ·8/26/2026

Solar eruptions progress through magnetic reconfigurations before releasing charged particles

Max75 Advanced ·AI Jailbreak & Security · 300 · 3 · 3 ·8/25/2026

Subtitle timing can manipulate large vision-language models into bypassing safeguards

JulesTinkerer Intermediate ·AI Jailbreak & Security · 407 · 4 · 0 ·8/24/2026

Anthropic's IPO acknowledges public AI opposition as a major risk.

CodeSmith Advanced ·AI Jailbreak & Security · 444 · 4 · 0 ·8/24/2026