# Evidence-Consistent Generative Detection under Scenario-Level Distribution Shift

arXiv: https://arxiv.org/abs/2608.21043 · HTML: https://mlsanity.com/paper/2608.21043/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 22 of its 44 references that are in the map.
- First posted on arXiv 21 Aug 2026.

## Builds on (most cited first, 22 of 22)

- [DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter](https://mlsanity.com/paper/1910.01108/) (2019)
- [Robust Speech Recognition via Large-Scale Weak Supervision](https://mlsanity.com/paper/2212.04356/) (2022)
- [Shortcut Learning in Deep Neural Networks](https://mlsanity.com/paper/2004.07780/) (2020)
- [Beyond Accuracy: Behavioral Testing of NLP models with CheckList](https://mlsanity.com/paper/2005.04118/) (2020)
- [Right for the Wrong Reasons: Diagnosing Syntactic Heuristics in Natural Language Inference](https://mlsanity.com/paper/1902.01007/) (2019)
- [Towards Faithfully Interpretable NLP Systems: How should we define and evaluate faithfulness?](https://mlsanity.com/paper/2004.03685/) (2020)
- [ERASER: A Benchmark to Evaluate Rationalized NLP Models](https://mlsanity.com/paper/1911.03429/) (2019)
- [e-SNLI: Natural Language Inference with Natural Language Explanations](https://mlsanity.com/paper/1812.01193/) (2018)
- [Learning the Difference that Makes a Difference with Counterfactually-Augmented Data](https://mlsanity.com/paper/1909.12434/) (2019)
- [Pretrained Transformers Improve Out-of-Distribution Robustness](https://mlsanity.com/paper/2004.06100/) (2020)
- [Measuring Association Between Labels and Free-Text Rationales](https://mlsanity.com/paper/2010.12762/) (2020)
- [Generating Fact Checking Explanations](https://mlsanity.com/paper/2004.05773/) (2020)
- [Faithfulness Tests for Natural Language Explanations](https://mlsanity.com/paper/2305.18029/) (2023)
- [SCOTT: Self-Consistent Chain-of-Thought Distillation](https://mlsanity.com/paper/2305.01879/) (2023)
- [State-of-the-art generalisation research in NLP: A taxonomy and review](https://mlsanity.com/paper/2210.03050/) (2022)
- [NILE : Natural Language Inference with Faithful Natural Language Explanations](https://mlsanity.com/paper/2005.12116/) (2020)
- [Leakage-Adjusted Simulatability: Can Models Generate Non-Trivial Explanations of Their Behavior in Natural Language?](https://mlsanity.com/paper/2010.04119/) (2020)
- [Types of Out-of-Distribution Texts and How to Detect Them](https://mlsanity.com/paper/2109.06827/) (2021)
- [PINTO: Faithful Language Reasoning Using Prompt-Generated Rationales](https://mlsanity.com/paper/2211.01562/) (2022)
- [Self-Training Meets Consistency: Improving LLMs' Reasoning with Consistency-Driven Rationale Evaluation](https://mlsanity.com/paper/2411.06387/) (2024)
- [Few Shot Rationale Generation using Self-Training with Dual Teachers](https://mlsanity.com/paper/2306.03315/) (2023)
- [Yesterday's News: Benchmarking Multi-Dimensional Out-of-Distribution Generalization of Misinformation Detection Models](https://mlsanity.com/paper/2410.18122/) (2024)

## Abstract (arXiv)

Conventional in-distribution evaluation can overestimate robustness when training and test data share recurring task-specific patterns or surface cues. This risk is especially relevant in social-engineering fraud detection, where attackers can preserve malicious intent while changing the scenario, impersonated entity, or wording. We study this problem as scenario-level out-of-distribution (SL-OOD) detection for SMS and voice phishing, where entire attack scenarios are held out from training while the label space remains fixed. This setting tests whether models can generalize to unseen attack scenarios using decision-relevant evidence rather than familiar scenario-specific cues. Using this SL-OOD evaluation, we find that high in-distribution performance does not reliably predict held-out robustness across feature-, encoder-, and decoder-based baselines. We interpret this gap as scenario memorization: reliance on recurring scenario-specific lexical or entity cues rather than decision-relevant evidence. We propose ECoG, an evidence-consistent generative framework that combines evidence-span supervision with a rationale-label consistency objective during training. On the 0.5B decoder, relative to the same backbone trained without consistency regularization, ECoG raises Macro-F1 on OOD challenging instances by 3.22 points, reduces the share of predictions whose generated rationale supports the opposite label by 4.22 points, and increases token-level overlap with reference evidence spans by 8.38 points; the reduction in prediction-rationale inconsistency is consistent across four decoder backbones. These results suggest that compact generative detectors can benefit from evidence supervision and rationale-label consistency under social-engineering shift.