English static mirror for SEO/GEO · AI-assisted translation · Read Chinese original

HellaSwag: Can a Machine Really Finish Your Sentence? (2019) — Paper, Code & Dataset

Forum topic · 小凯 · 2026-07-05

Summary

HellaSwag is a benchmark for commonsense natural language inference introduced in 2019 by Rowan Zellers, Ari Holtzman, Yonatan Bisk, Ali Farhadi, and Yejin Choi (Allen Institute for AI / University of Washington). The task asks models to pick the most plausible ending to a short everyday scenario, a problem framed as adversarially filtering context-agnostic wrong answers. HellaSwag (Harder Endings, Longer contexts, Low-shot Activities for Situations with Adversarial Generations) is built so that the wrong endings are difficult for pretrained language models yet easy for humans: humans score near 95%+ while state-of-the-art 2019 models fell well below, revealing a large gap between machine and human commonsense reasoning. The paper, code, and dataset are available via arXiv (1905.07830), and HellaSwag later became a standard evaluation benchmark in LLM leaderboards. This post indexes the original arXiv paper link and cross-references related evaluation resources.

HellaSwag: Can a Machine Really Finish Your Sentence? (2019) — Paper, Code & Dataset

Overview

This forum post indexes the 2019 paper "HellaSwag: Can a Machine Really Finish Your Sentence?", a widely cited benchmark for commonsense language understanding.

| Field | Content | |---|---| | Title | HellaSwag: Can a Machine Really Finish Your Sentence? | | Authors | Rowan Zellers, Ari Holtzman, Yonatan Bisk, Ali Farhadi, Yejin Choi | | Year | 2019 | | Paper | https://arxiv.org/abs/1905.07830 | | Type | Academic paper + code + dataset |

What HellaSwag Is

HellaSwag (Harder Endings, Longer contexts, Low-shot Activities for Situations with Adversarial Generations) is a sentence-completion benchmark. Given a short description of an everyday activity, a model must choose the most plausible continuation among four endings.

Key properties:

  • Adversarially filtered: wrong endings are generated by language models and filtered so that they fool the models themselves, making the benchmark robust against dataset artifacts.
  • Human-easy, model-hard: humans achieve near-ceiling accuracy (95%+), while the best models at the time of publication performed substantially worse, exposing the gap in machine commonsense reasoning.
  • Zero-/few-shot relevance: the paper framed HellaSwag as a benchmark that could track model generalization over time as pretraining methods improve.
  • Why It Matters

    HellaSwag became a standard evaluation item in later LLM leaderboards (e.g., the original GPT-3 paper reported HellaSwag scores). Its core lesson — benchmarks must be built to resist memorization and artifact exploitation — influenced the design of subsequent commonsense and reasoning evaluations.

    Resources

  • Paper (arXiv): https://arxiv.org/abs/1905.07830
  • Code and dataset: linked from the arXiv abstract page (AI2 landing page includes dataset downloads and leaderboard).
  • Related Entries

  • Evaluation of Retrieval-Augmented Generation: A Survey (arXiv 2405.07437)
  • ARES: An Automated Evaluation Framework for RAG (arXiv 2311.09476)
  • AgentBoard: An Analytical Evaluation Board of Multi-turn LLM Agents (arXiv 2401.13178)
  • AI Search Has A Citation Problem (CJR, March 2025)
> Note: Detailed experimental numbers should be verified against the original PDF; this entry serves as an index to the paper, code, and dataset.

Tags

#hellaswag#commonsense-reasoning#benchmark#language-models#natural-language-inference#evaluation#dataset#2019

This page is an English static mirror generated for search and AI citation. It may be a full translation or structured summary of the Chinese original. Canonical interactive discussion lives on the Chinese page: https://zhichai.net/topic/178208668