# When Choosing Plausible Alternatives, Clever Hans can be Clever

arXiv: https://arxiv.org/abs/1911.00225 · HTML: https://mlsanity.com/paper/1911.00225/

## Key facts

- Cited by 40 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 1 Nov 2019.

## Built on by (most cited first, 30 of 40)

- [Clever Hans or Neural Theory of Mind? Stress Testing Social Reasoning in Large Language Models](https://mlsanity.com/paper/2305.14763/) (2023)
- [SuperBPE: Space Travel for Language Models](https://mlsanity.com/paper/2503.13423/) (2025)
- [Mind the instructions: a holistic evaluation of consistency and interactions in prompt-based learning](https://mlsanity.com/paper/2310.13486/) (2023)
- [Language Models Represent Beliefs of Self and Others](https://mlsanity.com/paper/2402.18496/) (2024)
- [It Takes Two to Tango: Navigating Conceptualizations of NLP Tasks and Measurements of Performance](https://mlsanity.com/paper/2305.09022/) (2023)
- [COLA: Contextualized Commonsense Causal Reasoning from the Causal Inference Perspective](https://mlsanity.com/paper/2305.05191/) (2023)
- [FaithLM: Towards Faithful Explanations for Large Language Models](https://mlsanity.com/paper/2402.04678/) (2024)
- [SEA-HELM: Southeast Asian Holistic Evaluation of Language Models](https://mlsanity.com/paper/2502.14301/) (2025)
- [ROCK: Causal Inference Principles for Reasoning about Commonsense Causality](https://mlsanity.com/paper/2202.00436/) (2022)
- [CausalBERT: Injecting Causal Knowledge Into Pre-trained Models with Minimal Supervision](https://mlsanity.com/paper/2107.09852/) (2021)
- [Are Prompt-based Models Clueless?](https://mlsanity.com/paper/2205.09295/) (2022)
- [Are LLMs Socially Adaptive? Contrasting Belief Evolution in Large Language Models and Humans](https://mlsanity.com/paper/2410.10398/) (2024)
- [mCSQA: Multilingual Commonsense Reasoning Dataset with Unified Creation Strategy by Language Models and Humans](https://mlsanity.com/paper/2406.04215/) (2024)
- [Improving Commonsense Causal Reasoning by Adversarial Training and Data Augmentation](https://mlsanity.com/paper/2101.04966/) (2021)
- [Leaving the barn door open for Clever Hans: Simple features predict LLM benchmark answers](https://mlsanity.com/paper/2410.11672/) (2024)
- [COPA-SSE: Semi-structured Explanations for Commonsense Reasoning](https://mlsanity.com/paper/2201.06777/) (2022)
- [Knowledge-Augmented Language Models for Cause-Effect Relation Classification](https://mlsanity.com/paper/2112.08615/) (2021)
- [CAPO: Cost-Aware Prompt Optimization](https://mlsanity.com/paper/2504.16005/) (2025)
- [Doing Good or Doing Right? Exploring the Weakness of Commonsense Causal Reasoning Models](https://mlsanity.com/paper/2107.01791/) (2021)
- [Average Is Not Enough: Caveats of Multilingual Evaluation](https://mlsanity.com/paper/2301.01269/) (2023)
- [Batayan: A Filipino NLP benchmark for evaluating Large Language Models](https://mlsanity.com/paper/2502.14911/) (2025)
- [Learning to Learn to be Right for the Right Reasons](https://mlsanity.com/paper/2104.11514/) (2021)
- [Improving the OOD Performance of Closed-Source LLMs on NLI Through Strategic Data Selection](https://mlsanity.com/paper/2505.20209/) (2025)
- [NITP: Next Implicit Token Prediction for LLM Pre-training](https://mlsanity.com/paper/2605.24956/) (2026)
- [Generating Hypothetical Events for Abductive Inference](https://mlsanity.com/paper/2106.03973/) (2021)
- [AgentCoMa: A Compositional Benchmark Mixing Commonsense and Mathematical Reasoning in Real-World Scenarios](https://mlsanity.com/paper/2508.19988/) (2025)
- [Text-To-KG Alignment: Comparing Current Methods on Classification Tasks](https://mlsanity.com/paper/2306.02871/) (2023)
- [GLS-CSC: A Simple but Effective Strategy to Mitigate Chinese STM Models' Over-Reliance on Superficial Clue](https://mlsanity.com/paper/2309.04162/) (2023)
- [INSPECT: Intrinsic and Systematic Probing Evaluation for Code Transformers](https://mlsanity.com/paper/2312.05092/) (2023)
- [Out-of-Distribution Generalization in Text Classification: Past, Present, and Future](https://mlsanity.com/paper/2305.14104/) (2023)

## Abstract (arXiv)

Pretrained language models, such as BERT and RoBERTa, have shown large improvements in the commonsense reasoning benchmark COPA. However, recent work found that many improvements in benchmarks of natural language understanding are not due to models learning the task, but due to their increasing ability to exploit superficial cues, such as tokens that occur more often in the correct answer than the wrong one. Are BERT's and RoBERTa's good performance on COPA also caused by this? We find superficial cues in COPA, as well as evidence that BERT exploits these cues. To remedy this problem, we introduce Balanced COPA, an extension of COPA that does not suffer from easy-to-exploit single token cues. We analyze BERT's and RoBERTa's performance on original and Balanced COPA, finding that BERT relies on superficial cues when they are present, but still achieves comparable performance once they are made ineffective, suggesting that BERT learns the task to a certain degree when forced to. In contrast, RoBERTa does not appear to rely on superficial cues.