SAGE uses sparse autoencoders to boost vulnerability signals in LLMs, raising internal SNR 12.7x and delivering up to 318% MCC gains on vulnerability detection benchmarks.
Title resolution pending
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
citation-role summary
background 2
citation-polarity summary
years
2026 2verdicts
UNVERDICTED 2roles
background 2polarities
background 2representative citing papers
SpecRL uses the fraction of negative tests rejected by candidate specifications as a reward signal in RL training to produce stronger and more verifiable formal specifications than prior methods.
citing papers explorer
-
SAGE: Signal-Amplified Guided Embeddings for LLM-based Vulnerability Detection
SAGE uses sparse autoencoders to boost vulnerability signals in LLMs, raising internal SNR 12.7x and delivering up to 318% MCC gains on vulnerability detection benchmarks.
-
Reinforcement Learning with Negative Tests as Completeness Signal for Formal Specification Synthesis
SpecRL uses the fraction of negative tests rejected by candidate specifications as a reward signal in RL training to produce stronger and more verifiable formal specifications than prior methods.