Feature Attribution Stability Suite: How Stable Are Post-Hoc Attributions?
arXiv:2604.02532v1 Announce Type: new Abstract: Post-hoc feature attribution methods are widely deployed in safety-critical vision systems, yet their stability under realistic input perturbations remains poorly characterized. Existing metrics evaluate explanations primarily under additive noise, collapse stability to a single scalar, and fail to condition on prediction preservation, conflating explanation fragility with model sensitivity. We introduce the Feature Attribution Stability Suite (FASS), a benchmark that enforces prediction-invariance filtering, decomposes stability into three complementary metrics: structural similarity, rank correlation, and top-k Jaccard overlap-and evaluates across geometric, photometric, and compression perturbations. Evaluating four attribution methods (In
View PDF HTML (experimental)
Abstract:Post-hoc feature attribution methods are widely deployed in safety-critical vision systems, yet their stability under realistic input perturbations remains poorly characterized. Existing metrics evaluate explanations primarily under additive noise, collapse stability to a single scalar, and fail to condition on prediction preservation, conflating explanation fragility with model sensitivity. We introduce the Feature Attribution Stability Suite (FASS), a benchmark that enforces prediction-invariance filtering, decomposes stability into three complementary metrics: structural similarity, rank correlation, and top-k Jaccard overlap-and evaluates across geometric, photometric, and compression perturbations. Evaluating four attribution methods (Integrated Gradients, GradientSHAP, Grad-CAM, LIME) across four architectures and three datasets-ImageNet-1K, MS COCO, and CIFAR-10, FASS shows that stability estimates depend critically on perturbation family and prediction-invariance filtering. Geometric perturbations expose substantially greater attribution instability than photometric changes, and without conditioning on prediction preservation, up to 99% of evaluated pairs involve changed predictions. Under this controlled evaluation, we observe consistent method-level trends, with Grad-CAM achieving the highest stability across datasets.
Comments: Accepted in the proceedings track of XAI4CV Workshop at CVPR 2026. It has 2 images, 5 tables, 6 equations, and 35 references in the main paper and 12 figures, 15 tables, and 3 references in the supplementary material
Subjects:
Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
Cite as: arXiv:2604.02532 [cs.CV]
(or arXiv:2604.02532v1 [cs.CV] for this version)
https://doi.org/10.48550/arXiv.2604.02532
arXiv-issued DOI via DataCite (pending registration)
Submission history
From: Kamalasankari Subramaniakuppusamy [view email] [v1] Thu, 2 Apr 2026 21:32:54 UTC (1,261 KB)
Sign in to highlight and annotate this article
Conversation starters
Daily AI Digest
Get the top 5 AI stories delivered to your inbox every morning.
More about
modelbenchmarkannounce
StoryScope: Investigating idiosyncrasies in AI fiction
arXiv:2604.03136v1 Announce Type: new Abstract: As AI-generated fiction becomes increasingly prevalent, questions of authorship and originality are becoming central to how written work is evaluated. While most existing work in this space focuses on identifying surface-level signatures of AI writing, we ask instead whether AI-generated stories can be distinguished from human ones without relying on stylistic signals, focusing on discourse-level narrative choices such as character agency and chronological discontinuity. We propose StoryScope, a pipeline that automatically induces a fine-grained, interpretable feature space of discourse-level narrative features across 10 dimensions. We apply StoryScope to a parallel corpus of 10,272 writing prompts, each written by a human author and five LLM
Domain-Adapted Retrieval for In-Context Annotation of Pedagogical Dialogue Acts
arXiv:2604.03127v1 Announce Type: new Abstract: Automated annotation of pedagogical dialogue is a high-stakes task where LLMs often fail without sufficient domain grounding. We present a domain-adapted RAG pipeline for tutoring move annotation. Rather than fine-tuning the generative model, we adapt retrieval by fine-tuning a lightweight embedding model on tutoring corpora and indexing dialogues at the utterance level to retrieve labeled few-shot demonstrations. Evaluated across two real tutoring dialogue datasets (TalkMoves and Eedi) and three LLM backbones (GPT-5.2, Claude Sonnet 4.6, Qwen3-32b), our best configuration achieves Cohen's $\kappa$ of 0.526-0.580 on TalkMoves and 0.659-0.743 on Eedi, substantially outperforming no-retrieval baselines ($\kappa = 0.275$-$0.413$ and $0.160$-$0.4
Multi-Aspect Knowledge Distillation for Language Model with Low-rank Factorization
arXiv:2604.03110v1 Announce Type: new Abstract: Knowledge distillation is an effective technique for pre-trained language model compression. However, existing methods only focus on the knowledge distribution among layers, which may cause the loss of fine-grained information in the alignment process. To address this issue, we introduce the Multi-aspect Knowledge Distillation (MaKD) method, which mimics the self-attention and feed-forward modules in greater depth to capture rich language knowledge information at different aspects. Experimental results demonstrate that MaKD can achieve competitive performance compared with various strong baselines with the same storage parameter budget. In addition, our method also performs well in distilling auto-regressive architecture models.
Knowledge Map
Connected Articles — Knowledge Graph
This article is connected to other articles through shared AI topics and tags.
More in Models
StoryScope: Investigating idiosyncrasies in AI fiction
arXiv:2604.03136v1 Announce Type: new Abstract: As AI-generated fiction becomes increasingly prevalent, questions of authorship and originality are becoming central to how written work is evaluated. While most existing work in this space focuses on identifying surface-level signatures of AI writing, we ask instead whether AI-generated stories can be distinguished from human ones without relying on stylistic signals, focusing on discourse-level narrative choices such as character agency and chronological discontinuity. We propose StoryScope, a pipeline that automatically induces a fine-grained, interpretable feature space of discourse-level narrative features across 10 dimensions. We apply StoryScope to a parallel corpus of 10,272 writing prompts, each written by a human author and five LLM
Domain-Adapted Retrieval for In-Context Annotation of Pedagogical Dialogue Acts
arXiv:2604.03127v1 Announce Type: new Abstract: Automated annotation of pedagogical dialogue is a high-stakes task where LLMs often fail without sufficient domain grounding. We present a domain-adapted RAG pipeline for tutoring move annotation. Rather than fine-tuning the generative model, we adapt retrieval by fine-tuning a lightweight embedding model on tutoring corpora and indexing dialogues at the utterance level to retrieve labeled few-shot demonstrations. Evaluated across two real tutoring dialogue datasets (TalkMoves and Eedi) and three LLM backbones (GPT-5.2, Claude Sonnet 4.6, Qwen3-32b), our best configuration achieves Cohen's $\kappa$ of 0.526-0.580 on TalkMoves and 0.659-0.743 on Eedi, substantially outperforming no-retrieval baselines ($\kappa = 0.275$-$0.413$ and $0.160$-$0.4
Multi-Aspect Knowledge Distillation for Language Model with Low-rank Factorization
arXiv:2604.03110v1 Announce Type: new Abstract: Knowledge distillation is an effective technique for pre-trained language model compression. However, existing methods only focus on the knowledge distribution among layers, which may cause the loss of fine-grained information in the alignment process. To address this issue, we introduce the Multi-aspect Knowledge Distillation (MaKD) method, which mimics the self-attention and feed-forward modules in greater depth to capture rich language knowledge information at different aspects. Experimental results demonstrate that MaKD can achieve competitive performance compared with various strong baselines with the same storage parameter budget. In addition, our method also performs well in distilling auto-regressive architecture models.
Discussion
Sign in to join the discussion
No comments yet — be the first to share your thoughts!