Pushing the Boundaries of Multiple Choice Evaluation to One Hundred Options

TL;DR

This paper introduces a large-scale multiple choice evaluation method with 100 options to better assess language model reliability and reveal hidden weaknesses not visible in traditional low-option tests.

Contribution

It proposes a novel evaluation protocol with 100 options, demonstrating its effectiveness in exposing model failures and biases that are hidden in standard benchmarks.

Findings

High performance in low option settings can overstate model competence.

Dense distractors reveal gaps in model understanding and biases.

Candidate ranking, not context length, is the main bottleneck.

Abstract

Multiple choice evaluation is widely used for benchmarking large language models, yet near ceiling accuracy in low option settings can be sustained by shortcut strategies that obscure true competence. Therefore, we propose a massive option evaluation protocol that scales the candidate set to one hundred options and sharply reduces the impact of chance performance. We apply this framework to a Korean orthography error detection task where models must pick the single incorrect sentence from a large candidate set. With fixed targets and repeated resampling and shuffling, we obtain stable estimates while separating content driven failures from positional artifacts. Across experiments, results indicate that strong performance in low option settings can overstate model competence. This apparent advantage often weakens under dense interference at high $N$, revealing gaps that conventional benchmarks tend to obscure. We identify two failure modes, semantic confusion and position bias toward early options under uncertainty. To isolate the effect of context length, we run padding controlled and length matched tests, which suggest that the main bottleneck is candidate ranking rather than context length. Together, these findings support massive option evaluation as a general framework for stress testing model reliability under extreme distractor density, beyond what low option benchmarks can reveal.