LLM Safety - Edgentiq

Unveiling LLM Refusal: A Multi-Directional Approach Using Self-Organizing Maps

Bridging Safety Gaps in Large Language Models with Policy Patches

The Fading Footprints: How Fine-Tuning Impacts Knowledge Edits in Language Models

Unveiling the Dual Nature of LLM Safety: A New Framework to Bypass Alignment

LEAP: A New Approach to Combat AI Hallucinations

spot_img

Recently Added

Enhancing LLM Safety: A Multi-Agent Debate Approach for Efficient Evaluation

Read more

MENTOR: Teaching LLMs to Self-Assess and Mitigate Hidden Dangers in Domain Tasks

Read more

AdversariaLLM: A New Framework for Consistent LLM Robustness Evaluation

Read more

Guiding LLMs to Safer Responses with Feature Steering

Read more

Advanced LLM Jailbreaking: Co-Evolving Prompts and Evaluation for Robustness

Read more

Enhancing LLM Safety Through Smart Data Selection During Fine-Tuning

Read more

Activating LLM’s Inner Guard: A New Defense Against Jailbreak Attacks

Read more

JT-Safe: Building Inherently Safer and More Trustworthy AI Through Context-Rich Data

Read more

Enhancing LLM Safety: A New Approach to Deep Alignment

Read more

Unmasking AI Jailbreaks: A New Framework for Detecting Adversarial Attacks on Language Models

Read more

Unmasking AI’s Dark Side: How LLMs Can Be Coerced into Multi-Turn Harassment

Read more

GuardSpace: A New Approach to Preserving AI Safety During Language Model Fine-Tuning

Read more

Safeguarding Enterprise AI: Protect’s Multi-Modal Approach to LLM Safety

Read more

Uncovering Hidden Dangers: A New Approach to Red-Teaming LLMs with Web Search

Read more

Navigating Mental Health AI: A Framework for Safer Disclosure and Enhanced User Understanding

Read more

LLM Safety Alert: Reduced Response Diversity Makes AI Reasoning Unsafe

Read more

Guiding Language Models with Energy: Fewer False Refusals, More Helpful Responses

Read more

Precision Forgetting: How SIMU Enhances LLM Unlearning by Targeting Critical Neurons

Read more

Securing LLMs Against Evolving Jailbreak Attacks with MetaDefense

Read more

VeriGuard: A New Framework for Ensuring Safety in AI Agents Through Verified Code

Read more

New Benchmark Reveals Language Model Vulnerabilities to Sociopolitical Harms

Read more

Assessing Catastrophic Risks in LLM Conversations

Read more

Pinpointing the Origins of Undesirable LLM Behaviors

Read more

Language Models Gain Self-Correction for Safer Outputs

Read more

Enhancing LLM Safety with Adaptive Upcycling and Temperature Control

Read more

RADAR: A New Framework for Enhanced LLM Safety Evaluation

Read more

ASGUARD: A Targeted Defense Against Tense-Based LLM Jailbreaks

Read more

SafeEvalAgent: A Dynamic Approach to AI Safety Evaluation

Read more

Unmasking the ‘Wrong Chat’ Problem: How Large Language Models Fail at Their Intended Purpose

Read more

Unmasking the Vulnerabilities of AI Search Agents: A New Framework for Safety Assessment

Read more

Gen AI News and Updates

spot_img

- Advertisement -

Unveiling LLM Refusal: A Multi-Directional Approach Using Self-Organizing Maps

November 12, 2025

Bridging Safety Gaps in Large Language Models with Policy Patches

November 12, 2025

The Fading Footprints: How Fine-Tuning Impacts Knowledge Edits in Language Models

November 11, 2025