A Human-in-the-Loop Corpus for LLM-Based Simplification of Scientific Summaries

Michael Färber, Kyuri Im

arXiv · PDF · Open in the Atlas

Code that ran

We lifted 2 functions out of this paper's own repositories and ran 1 of them in a sandbox. "Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results.

Repositories linked to this paper

• faerber-lab/scientific-text-simplification-corpus

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

Interdisciplinary research is accelerating, yet scientific papers remain difficult to read outside their home fields. We study large language model (LLM)-based simplification for scientific texts and present a human-in-the-loop workflow that turns expert summaries into more accessible versions for non-specialists. Using SciSummNet as the source corpus, we first generate baseline simplifications with GPT-4o-mini. In Phase 1, readers from STEM domains outside computer science identify difficult sentences and phrases and compare the original and GPT-simplified summaries in terms of understanding, naturalness, and simplicity. In Phase 2, computer science experts use this feedback to produce expert-edited reference simplifications. We release the resulting corpus together with human judgments and automatic evaluation results. The Phase 1 judgments show a clear preference for the GPT outputs in understanding and simplicity, while qualitative analysis of the Phase 2 edits illustrates the importance of preserving domain terminology and scientific claim strength. The resource supports the training and benchmarking of simplification systems for cross-disciplinary scientific communication. Dataset: https://github.com/faerber-lab/scientific-text-simplification-corpus

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2607.25630")
get_code_for_paper("2607.25630")
have("2607.25630")

Connect an agent — have() is free.