Syn-QA2: Evaluating False Assumptions in Long-tail Questions with Synthetic QA Datasets
This work addresses a gap in robust QA evaluation for researchers and practitioners, but it is incremental as it builds on prior work by extending analysis to synthetic long-tail data.
The paper tackled the problem of evaluating QA systems' sensitivity to false assumptions in long-tail questions by introducing Syn-QA2, a synthetic dataset, and found that detection tasks are particularly challenging, especially for long-tail questions, with models showing low performance.
Sensitivity to false assumptions (or false premises) in information-seeking questions is critical for robust question-answering (QA) systems. Recent work has shown that false assumptions in naturally occurring questions pose challenges to current models, with low performance on both generative QA and simple detection tasks (Kim et al. 2023). However, the focus of existing work on naturally occurring questions leads to a gap in the analysis of model behavior on the long tail of the distribution of possible questions. To this end, we introduce Syn-(QA)$^2$, a set of two synthetically generated QA datasets: one generated using perturbed relations from Wikidata, and the other by perturbing HotpotQA (Yang et al. 2018). Our findings from evaluating a range of large language models are threefold: (1) false assumptions in QA are challenging, echoing the findings of prior work, (2) the binary detection task is challenging even compared to the difficulty of generative QA itself, possibly due to the linguistic structure of the problem, and (3) the detection task is more challenging with long-tail questions compared to naturally occurring questions, highlighting the utility of our synthetic datasets and generation method.