post-hoc-analysis

This paper shows that the geometric symmetry visible from neural network weights depends on the positional encoding and readout observable, and validates this using MLPs trained on 2D signed distance functions with multiple symmetry groups.

MemAudit is a post-hoc auditing framework for memory-augmented LLM agents that identifies poisoned memories by combining counterfactual influence scores and structural anomaly detection, reducing attack success rates from over 70% to 0% in realistic scenarios.