Mova0.5 pre-release⌕
☰
home / docs / papers / authors / Hannaneh Hajishirzi

Hannaneh Hajishirzi

3 paper cards in Mova's library · 2014–2023

Hannaneh Hajishirzi wrote 3 of the papers Mova reads, in 2014–2023; often with Ally Zhang, Ananth Gottumukkala, Ben Bogin.

OpenAlex · ORCID · h-index 63 · 18822 citations · 287 works — openalex, 2026-10-18

The profile is the library's match of the name; two people with one name are not always told apart.

Elsewhere: Google Scholar · Semantic Scholar · dblp · ACL Anthology · ORCID

Papers

  1. CREPE: Open-Domain Question Answering with False Presuppositions (2023)
    CREPE has 8,400 «explain like I'm five» questions from Reddit. Annotators, guided by the most upvoted answer, mark whether the question rests on a false presupposition and, if so, write the presupposition and a correction. About 25% of real information-seeking questions contain one. A sample of 50 shows the kinds: a…
  2. Evaluating Models' Local Decision Boundaries via Contrast Sets (2020) measured
    Standard test sets come from the same distribution as the training data, so a dataset with systematic gaps or annotation artifacts lets models score well with shallow heuristics.
  3. Learning to Solve Arithmetic Word Problems with Verb Categorization (2014)
    The paper presents ARIS, an early system for addition and subtraction word problems that does not depend on equation templates.

Co-authors (31)

This page in Markdown → https://md.mova.elamur.ai/free/en/docs/papers/authors/hannaneh-hajishirzi/