7 papers
CIPHER: Cryptographic Insecurity Profiling via Hybrid Evaluation of Responses
Max Manolov, Tony Gao, Siddharth Shukla +2
Large language models (LLMs) are increasingly used to assist developers with code, yet their implementations of cryptographic functionality often contain exploitable flaws. Minor d…
A Few Bad Neurons: Isolating and Surgically Correcting Sycophancy
Claire O'Brien, Jessica Seto, Dristi Roy +6
Behavioral alignment in large language models (LLMs) is often achieved through broad fine-tuning, which can result in undesired side effects like distributional shift and low inter…
Chopping Trees: Semantic Similarity Based Dynamic Pruning for Tree-of-Thought Reasoning
Joongho Kim, Xirui Huang, Zarreen Reza +1
Tree-of-Thought (ToT) reasoning boosts the problem-solving abilities of Large Language Models (LLMs) but is computationally expensive due to semantic redundancy, where distinct bra…
Alignment-Constrained Dynamic Pruning for LLMs: Identifying and Preserving Alignment-Critical Circuits
Dev Patel, Gabrielle Gervacio, Diekola Raimi +5
Large Language Models require substantial computational resources for inference, posing deployment challenges. While dynamic pruning offers superior efficiency over static methods…
Discovering Transformer Circuits via a Hybrid Attribution and Pruning Framework
Hao Gu, Vibhas Nair, Amrithaa Ashok Kumar +2
Interpreting language models often involves circuit analysis, which aims to identify sparse subnetworks, or circuits, that accomplish specific tasks. Existing circuit discovery alg…
Evaluation Awareness Scales Predictably in Open-Weights Large Language Models
Maheep Chaudhary, Ian Su, Nikhil Hooda +6
Large language models (LLMs) can internally distinguish between evaluation and deployment contexts, a behaviour known as \emph{evaluation awareness}. This undermines AI safety eval…