Papers with Code 아카이브

Red Teaming

태스크의 하위 분류와 논문·코드·평가·기법·데이터셋을 탐색하세요.

전체 기법전체 데이터셋

하위 태스크

0

평가표

1

보존된 평가 기록

1

코드

관련 논문과 평가 결과에 연결된 구현입니다. 출처에서 확인된 저자 구현에는 공식 코드 표시가 있습니다.

https://github.com/paul-rottger/exaggerated-safety

공식 코드

XSTest: A Test Suite for Identifying Exaggerated Safety Behaviours in Large Language Models

https://github.com/princeton-sysml/jailbreak_llm

공식 코드

Catastrophic Jailbreak of Open-source LLMs via Exploiting Generation

https://github.com/redteamingforllms/redteamingforllms

공식 코드

Red Teaming for Large Language Models At Scale: Tackling Hallucinations on Mathematics Tasks

https://github.com/rgreenblatt/control-evaluations

공식 코드

AI Control: Improving Safety Despite Intentional Subversion

https://github.com/richterleo/Auditing_Test_for_LMs

공식 코드

An Auditing Test To Detect Behavioral Shift in Language Models

https://github.com/rishub-tamirisa/tamper-resistance

공식 코드

Tamper-Resistant Safeguards for Open-Weight LLMs

https://github.com/robustnlp/cipherchat

공식 코드

GPT-4 Is Too Smart To Be Safe: Stealthy Chat with LLMs via Cipher

https://github.com/rucnyz/redagent

공식 코드

PrivAgent: Agentic-based Red-teaming for LLM Privacy Leakage

https://github.com/safolab-wisc/autodan-turbo

공식 코드

AutoDAN-Turbo: A Lifelong Agent for Strategy Self-Exploration to Jailbreak LLMs

https://github.com/sail-sg/agent-smith

공식 코드

Agent Smith: A Single Image Can Jailbreak One Million Multimodal LLM Agents Exponentially Fast

https://github.com/scaleapi/browser-art

공식 코드

Refusal-Trained LLMs Are Easily Jailbroken As Browser Agents

https://github.com/scimt/scimt-benchmark

공식 코드

Control Risk for Potential Misuse of Artificial Intelligence in Science

https://github.com/sherdencooper/gptfuzz

공식 코드

GPTFUZZER: Red Teaming Large Language Models with Auto-Generated Jailbreak Prompts

https://github.com/shizhouxing/Attack-LM-Detectors

공식 코드

Red Teaming Language Model Detectors with Language Models

https://github.com/silent-zebra/twisted-smc-lm

공식 코드

Probabilistic Inference in Language Models via Twisted Sequential Monte Carlo

https://github.com/sisl/astprompter

공식 코드

ASTPrompter: Weakly Supervised Automated Language Model Red-Teaming to Identify Low-Perplexity Toxic Prompts

https://github.com/sisl/kov.jl

공식 코드

Kov: Transferable and Naturalistic Black-Box LLM Attacks using Markov Decision Processes and Tree Search

https://github.com/snu-mllab/bayesian-red-teaming

공식 코드

Query-Efficient Black-Box Red Teaming via Bayesian Optimization

https://github.com/sqzhang-lazy/guidelinellm

공식 코드

Look Before You Leap: Enhancing Attention and Vigilance Regarding Harmful Content with GuidelineLLM

https://github.com/sufenlp/AttanttionShiftJailbreak

공식 코드

Distract Large Language Models for Automatic Jailbreak Attack

https://github.com/techsachinkr/Overriding_Model_Safety_Protections

공식 코드

Overriding Safety protections of Open-source Models

https://github.com/thestephencasper/common_claim

공식 코드

Explore, Establish, Exploit: Red Teaming Language Models from Scratch

https://github.com/thestephencasper/explore_establish_exploit_llms

공식 코드

Explore, Establish, Exploit: Red Teaming Language Models from Scratch

https://github.com/thestephencasper/latent_adversarial_training

공식 코드

Defending Against Unforeseen Failure Modes with Latent Adversarial Training

https://github.com/thestephencasper/lm_white_box_attacks

공식 코드

Red Teaming with Mind Reading: White-Box Adversarial Policies Against RL Agents

https://github.com/thestephencasper/white_box_rarl

공식 코드

Red Teaming with Mind Reading: White-Box Adversarial Policies Against RL Agents

https://github.com/tjunlp-lab/aprt

공식 코드

Automated Progressive Red Teaming

https://github.com/tmllab/2025_ICLR_PiF

공식 코드

Understanding and Enhancing the Transferability of Jailbreaking Attacks

https://github.com/wang2226/backdoor-activation-attack

공식 코드

Trojan Activation Attack: Red-Teaming Large Language Models using Activation Steering for Safety-Alignment

https://github.com/x-zheng16/redrft

공식 코드

RedRFT: A Light-Weight Benchmark for Reinforcement Fine-Tuning-Based Red Teaming

https://github.com/yang-yan-yang-yan/sop

공식 코드

SeqAR: Jailbreak LLMs with Sequential Auto-Generated Characters

https://github.com/yiyiyizhao/siren

공식 코드

Siren: A Learning-Based Multi-Turn Attack Framework for Simulating Real-World Human Jailbreak Behaviors

https://github.com/yuki-younai/mtsa

공식 코드

MTSA: Multi-turn Safety Alignment for LLMs through Multi-round Red-teaming

https://github.com/zhxieml/remiss-jailbreak

공식 코드

Jailbreaking as a Reward Misspecification Problem

https://huggingface.co/datasets/SoftMINER-Group/NicheHazardQA

공식 코드

Sowing the Wind, Reaping the Whirlwind: The Impact of Editing Language Models

https://github.com/AIM-Intelligence/Automated-Multi-Turn-Jailbreaks

LLM Defenses Are Not Robust to Multi-Turn Human Jailbreaks Yet

출처: Papers with Code · CC BY-SA 4.0. 아카이브 자료와 새로 수집한 논문은 별도로 관리합니다.