[NetSec] Understanding TLS, PKI, and the Post-Quantum Transition
An amateur analysis of two TLS sessions, from cipher suites and certificate chains to Kyber/ML-KEM migration(ENJOY!!).
C2SR Lab.
[PaRev] INFA-Guard: Infection-Aware MAS Safeguarding
Infection-aware defense for malicious propagation in LLM multi-agent systems.
[PaRev] NDAD: Negative-Direction Aware Decoding for Large Language Models via Controllable Hallucination Signal Injection
A review of NDAD, a training-free decoding method that suppresses hallucination by constructing and subtracting controllable hallucination signals.
![[ICLR 2025] On the Role of Attention Heads in Large Language Model Safety](/assets/img/posts/Ships_Sahara/thumbnail.png)
[PaRev] On the Role of Attention Heads in Large Language Models Safety
Attention heads play a critical role in shaping both safe and harmful behaviors
[PaRev] Goal-Aware Identification and Rectification of Misinformation in Multi-Agent Systems
A review of ARGUS, a training-free framework for locating and rectifying misinformation flows in LLM-based multi-agent systems.
[PaRev] G-Safeguard: A Topology-Guided Security Lens for LLM-based Multi-Agent Systems
A review of G-Safeguard, a graph-based framework for detecting and remediating attacks in LLM-based multi-agent systems.
[PaRev] Programming Refusal with Conditional Activation Steering (CAST)
A review of CAST, an activation steering method for selectively applying refusal behavior without fine-tuning.
[cAIRev] Several Transformer Models
Several transformer models descriptions focus on its architecture
[CCE 2025] PaperLibrary write-up
The very first posting of C2SR Lab blog!!
[AISec] OWASP Top 10 for LLM Applications 2025
A security review of the OWASP Top 10 for LLM Applications 2025, with practical analysis for GenAI application design and threat modeling.