MirageBackdoor is the first backdoor attack that preserves clean chain-of-thought reasoning in LLMs while steering the final answer to a specific incorrect target under a trigger.
hub
Badchain: Backdoor chain-of-thought prompting for large language models.arXiv preprint arXiv:2401.12242
12 Pith papers cite this work. Polarity classification is still indexing.
12
Pith papers citing it
hub tools
citation-role summary
background 2
citation-polarity summary
roles
background 2polarities
background 2representative citing papers
Compilation optimizations can be exploited to create stealthy backdoors in LLMs that remain dormant without optimization but achieve ~90% attack success while preserving clean accuracy near 100%.
R-CoT embeds watermarks into LLM reasoning paths via redundant CoT and GRPO-based dual optimization, maintaining over 95% true positive rate under fine-tuning and post-training changes.
TIGS detects backdoor-induced attention collapse in LLMs and applies content-aware tail-risk screening plus intrinsic geometric smoothing to suppress attacks while preserving normal performance.
BadStyle creates stealthy backdoors in LLMs by poisoning samples with imperceptible style triggers and using an auxiliary loss to stabilize payload injection, achieving high attack success rates across multiple models while evading defenses.
An external zero-shot monitor detects nine unsafe reasoning behaviors in LLMs at 87% step-level accuracy with low false positives and low latency.
AttnTrace is an attention-weight-based context traceback method for LLMs that claims higher accuracy and efficiency than prior art like TracLLM while aiding prompt injection detection.
ShadowCoT introduces a reasoning-level backdoor attack on LLMs achieving 94.4% attack success rate and 88.4% hijacking success rate with 0.15% parameter updates via internal state conditioning and reasoning chain pollution.
Privacy attacks on LLMs show strong signals for membership inference and backdoors but weaker performance for attribute inference and data extraction, with risks highly dependent on system configuration.
A survey categorizing scaling in LLM reasoning across input size, steps, rounds, training, and future directions, noting that scaling can negatively affect performance.
Survey of harmful fine-tuning attacks on LLMs, their variants, defense strategies, mechanical analysis, and evaluation methodologies.
citing papers explorer
-
MirageBackdoor: A Stealthy Attack that Induces Think-Well-Answer-Wrong Reasoning
MirageBackdoor is the first backdoor attack that preserves clean chain-of-thought reasoning in LLMs while steering the final answer to a specific incorrect target under a trigger.
-
Trusted Weights, Treacherous Optimizations? Optimization-Triggered Backdoor Attacks on LLMs
Compilation optimizations can be exploited to create stealthy backdoors in LLMs that remain dormant without optimization but achieve ~90% attack success while preserving clean accuracy near 100%.
-
R-CoT: A Reasoning-Layer Watermark via Redundant Chain-of-Thought in Large Language Models
R-CoT embeds watermarks into LLM reasoning paths via redundant CoT and GRPO-based dual optimization, maintaining over 95% true positive rate under fine-tuning and post-training changes.
-
Defusing the Trigger: Plug-and-Play Defense for Backdoored LLMs via Tail-Risk Intrinsic Geometric Smoothing
TIGS detects backdoor-induced attention collapse in LLMs and applies content-aware tail-risk screening plus intrinsic geometric smoothing to suppress attacks while preserving normal performance.
-
Stealthy Backdoor Attacks against LLMs Based on Natural Style Triggers
BadStyle creates stealthy backdoors in LLMs by poisoning samples with imperceptible style triggers and using an auxiliary loss to stabilize payload injection, achieving high attack success rates across multiple models while evading defenses.
-
Beyond Content Safety: Real-Time Monitoring for Reasoning Vulnerabilities in Large Language Models
An external zero-shot monitor detects nine unsafe reasoning behaviors in LLMs at 87% step-level accuracy with low false positives and low latency.
-
AttnTrace: Contextual Attribution of Prompt Injection and Knowledge Corruption
AttnTrace is an attention-weight-based context traceback method for LLMs that claims higher accuracy and efficiency than prior art like TracLLM while aiding prompt injection detection.
-
ShadowCoT: Cognitive Hijacking for Stealthy Reasoning Backdoors in LLMs
ShadowCoT introduces a reasoning-level backdoor attack on LLMs achieving 94.4% attack success rate and 88.4% hijacking success rate with 0.15% parameter updates via internal state conditioning and reasoning chain pollution.
-
On the Privacy of LLMs: An Ablation Study
Privacy attacks on LLMs show strong signals for membership inference and backdoors but weaker performance for attribute inference and data extraction, with risks highly dependent on system configuration.
-
A Survey of Scaling in Large Language Model Reasoning
A survey categorizing scaling in LLM reasoning across input size, steps, rounds, training, and future directions, noting that scaling can negatively affect performance.
-
Harmful Fine-tuning Attacks and Defenses for Large Language Models: A Survey
Survey of harmful fine-tuning attacks on LLMs, their variants, defense strategies, mechanical analysis, and evaluation methodologies.
- SkillTrojan: Backdoor Attacks on Skill-Based Agent Systems