Reasoning Models are Test Exploiters: Rethinking Multiple-Choice

• URL: http://arxiv.org/abs/2507.15337v1
• Date: Mon, 21 Jul 2025 07:49:32 GMT
• Title: Reasoning Models are Test Exploiters: Rethinking Multiple-Choice
• Authors: Narun Raman, Taylor Lundy, Kevin Leyton-Brown,
• Abstract summary: Large Language Models (LLMs) are asked to choose among a fixed set of choices in question-answering domains.<n>Multiple-choice question-answering (McQCA) is a good proxy for the downstream performance of models as long as they were allowed to perform chain-of-thought reasoning.<n>We conclude that MCQA is no longer a good proxy for assessing downstream performance of state-of-the-art models.
• Score: 10.085788712670487
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: When evaluating Large Language Models (LLMs) in question-answering domains, it is common to ask the model to choose among a fixed set of choices (so-called multiple-choice question-answering, or MCQA). Although downstream tasks of interest typically do not provide systems with explicit options among which to choose, this approach is nevertheless widely used because it makes it makes automatic grading straightforward and has tended to produce challenging benchmarks that correlate sufficiently well with downstream performance. This paper investigates the extent to which this trend continues to hold for state-of-the-art reasoning models, describing a systematic evaluation of $15$ different question-answering benchmarks (e.g., MMLU, HLE) and $25$ different LLMs (including small models such as Qwen 7B and relatively large models such as Llama 70B). For each model-benchmark pair, we considered $5$ ways of presenting the model with questions, including variations on whether multiple choices were offered to the model at all; whether "none of the above" sometimes replaced the right answer; and whether the model was permitted to perform chain-of-thought reasoning before and/or after the choices were presented. MCQA remained a good proxy for the downstream performance of models as long as they were allowed to perform chain-of-thought reasoning only before being presented with the options among which they had to select. On the other hand, large models that were able to perform reasoning after being given a set of options tended to significantly outperform their free-text performance due to exploiting the information in the options. We conclude that MCQA is no longer a good proxy for assessing downstream performance of state-of-the-art models, and offer practical guidelines for designing more robust, bias-resistant benchmarks that better reflect LLMs' genuine reasoning capabilities.

Related papers

• Right Answer, Wrong Score: Uncovering the Inconsistencies of LLM Evaluation in Multiple-Choice Question Answering [78.89231943329885]
  Multiple-Choice Question Answering (MCQA) is widely used to evaluate Large Language Models (LLMs)<n>We show that multiple factors can significantly impact the reported performance of LLMs.<n>We analyze whether existing answer extraction methods are aligned with human judgment.
  arXiv  Detail & Related papers  (2025-03-19T08:45:03Z)
• Ranked from Within: Ranking Large Multimodal Models for Visual Question Answering Without Labels [64.94853276821992]
  Large multimodal models (LMMs) are increasingly deployed across diverse applications.<n>Traditional evaluation methods are largely dataset-centric, relying on fixed, labeled datasets and supervised metrics.<n>We explore unsupervised model ranking for LMMs by leveraging their uncertainty signals, such as softmax probabilities.
  arXiv  Detail & Related papers  (2024-12-09T13:05:43Z)
• Differentiating Choices via Commonality for Multiple-Choice Question Answering [54.04315943420376]
  Multiple-choice question answering can provide valuable clues for choosing the right answer. Existing models often rank each choice separately, overlooking the context provided by other choices. We propose a novel model by differentiating choices through identifying and eliminating their commonality, called DCQA.
  arXiv  Detail & Related papers  (2024-08-21T12:05:21Z)
• OLMES: A Standard for Language Model Evaluations [64.85905119836818]
  OLMES is a documented, practical, open standard for reproducible language model evaluations.<n>It supports meaningful comparisons between smaller base models that require the unnatural "cloze" formulation of multiple-choice questions.<n> OLMES includes well-considered, documented recommendations guided by results from existing literature as well as new experiments resolving open questions.
  arXiv  Detail & Related papers  (2024-06-12T17:37:09Z)
• A Study on Large Language Models' Limitations in Multiple-Choice Question Answering [0.0]
  We analyze 26 small open-source models and find that 65% of the models do not understand the task. Only 4 models properly select an answer from the given choices, and only 5 of these models are choice order independent.
  arXiv  Detail & Related papers  (2024-01-15T20:42:16Z)
• Causal Q-Aggregation for CATE Model Selection [20.39573621844382]
  We propose a new CATE ensembling approach based on Qaggregation using the doubly robust loss.<n>Our main result shows that causal Q-aggregation achieves statistically optimal model selection regret rates.
  arXiv  Detail & Related papers  (2023-10-25T19:27:05Z)
• HANS, are you clever? Clever Hans Effect Analysis of Neural Systems [1.6267479602370545]
  Large Language Models (It-LLMs) have been exhibiting outstanding abilities to reason around cognitive states, intentions, and reactions of all people involved, letting humans guide and comprehend day-to-day social interactions effectively. Several multiple-choice questions (MCQ) benchmarks have been proposed to construct solid assessments of the models' abilities. However, earlier works are demonstrating the presence of inherent "order bias" in It-LLMs, posing challenges to the appropriate evaluation.
  arXiv  Detail & Related papers  (2023-09-21T20:52:18Z)
• Large Language Models Are Not Robust Multiple Choice Selectors [117.72712117510953]
  Multiple choice questions (MCQs) serve as a common yet important task format in the evaluation of large language models (LLMs) This work shows that modern LLMs are vulnerable to option position changes due to their inherent "selection bias" We propose a label-free, inference-time debiasing method, called PriDe, which separates the model's prior bias for option IDs from the overall prediction distribution.
  arXiv  Detail & Related papers  (2023-09-07T17:44:56Z)
• Large Language Models Sensitivity to The Order of Options in Multiple-Choice Questions [5.187383020960245]
  Large Language Models (LLMs) have demonstrated remarkable capabilities in various NLP tasks. Previous works have shown these models are sensitive towards prompt wording, and few-shot demonstrations and their order. This paper investigates sensitivity of LLMs towards the order of options in multiple-choice questions.
  arXiv  Detail & Related papers  (2023-08-22T14:54:59Z)
• Getting MoRE out of Mixture of Language Model Reasoning Experts [71.61176122960464]
  We propose a Mixture-of-Reasoning-Experts (MoRE) framework that ensembles diverse specialized language models. We specialize the backbone language model with prompts optimized for different reasoning categories, including factual, multihop, mathematical, and commonsense reasoning. Our human study confirms that presenting expert predictions and the answer selection process helps annotators more accurately calibrate when to trust the system's output.
  arXiv  Detail & Related papers  (2023-05-24T02:00:51Z)
• Generative Context Pair Selection for Multi-hop Question Answering [60.74354009152721]
  We propose a generative context selection model for multi-hop question answering. Our proposed generative passage selection model has a better performance (4.9% higher than baseline) on adversarial held-out set.
  arXiv  Detail & Related papers  (2021-04-18T07:00:48Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.