post-hoc

Decoy Direction Optimization (DDO) is a fast, post-hoc defense method that protects open-weight LLMs from refusal feature ablation attacks by injecting decoy signals into the network, achieving high robustness at lower cost than trained defenses.

This paper introduces FALCON-Discover, a post-hoc framework for discovering regions where model predictions are confidently wrong, shifting calibration focus from aggregate metrics to sample-level dangerous failures.

Proposes SAGE, a post-hoc method to sanitize the final unlearning vector in LLMs, improving the retain-forget trade-off without rerunning the unlearning pipeline.

Proposes REEF-GP, a post-hoc uncertainty quantification framework that fits a Gaussian process to the residuals of a frozen neural operator using its internal embeddings, enabling geometry-aware and calibrated uncertainties at low cost.

A post-hoc method reduces spurious correlations in fine-tuned LLMs by truncating the tail of the SVD of the weight update matrix. It reduces the spurious-group gap by up to 5x with less than 2pp accuracy loss, without retraining or group labels.

Phase Marginalization is a post-hoc method that addresses phase-dependent instability in Vision Transformers by evaluating structured patch-grid phases and aggregating outputs. It improves segmentation, depth, and local matching over the canonical baseline with minimal extra cost.

CSULoRA is a post-hoc method for correcting trained LoRA adapters to preserve safety alignment while maintaining utility, using closest safe update estimation.