Pith. sign in

REVIEW 4 cited by

Hurdles to Progress in Long-form Question Answering

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2103.06332 v2 pith:VG262PM7 submitted 2021-03-10 cs.CL cs.LG

Hurdles to Progress in Long-form Question Answering

classification cs.CL cs.LG
keywords lfqaeli5progressquestionsystemansweransweringchallenges
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
read the original abstract

The task of long-form question answering (LFQA) involves retrieving documents relevant to a given question and using them to generate a paragraph-length answer. While many models have recently been proposed for LFQA, we show in this paper that the task formulation raises fundamental challenges regarding evaluation and dataset creation that currently preclude meaningful modeling progress. To demonstrate these challenges, we first design a new system that relies on sparse attention and contrastive retriever learning to achieve state-of-the-art performance on the ELI5 LFQA dataset. While our system tops the public leaderboard, a detailed analysis reveals several troubling trends: (1) our system's generated answers are not actually grounded in the documents that it retrieves; (2) ELI5 contains significant train / validation overlap, as at least 81% of ELI5 validation questions occur in paraphrased form in the training set; (3) ROUGE-L is not an informative metric of generated answer quality and can be easily gamed; and (4) human evaluations used for other text generation tasks are unreliable for LFQA. We offer suggestions to mitigate each of these issues, which we hope will lead to more rigorous LFQA research and meaningful progress in the future.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 4 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. RWGBench: Evaluating Scholarly Positioning in Related Work Generation

    cs.DL 2026-05 unverdicted novelty 7.0

    RWGBench is a citation-centric benchmark for related work generation built from 40k CS papers and a 100-paper test set, with multi-dimensional metrics that better match human expert judgment than standard similarity scores.

  2. RWGBench: Evaluating Scholarly Positioning in Related Work Generation

    cs.DL 2026-05 accept novelty 6.5

    RWGBench evaluates related-work generation as citation decision-making (selection, placement, organization, discourse) rather than text similarity, exposing retrieval and generation failures that standard metrics miss.

  3. RWGBench: Evaluating Scholarly Positioning in Related Work Generation

    cs.DL 2026-05 conditional novelty 6.0

    RWGBench measures related-work generation by citation choices, and shows citation-focused metrics expose failures that text-similarity and LLM-judge scores miss.

  4. RWGBench: Evaluating Scholarly Positioning in Related Work Generation

    cs.DL 2026-05 conditional novelty 5.5

    RWGBench evaluates related-work generation as citation decision-making with multi-dimensional metrics that track expert judgment better than ROUGE/BERTScore.