Paper Digest: 100 Must-Read Natural Language Processing Papers of the Past 10 Years (2016-2025)

Natural language processing has been rebuilt from the ground up in the past decade. Subword units made neural machine translation practical. ELMo, and then BERT, made pretrained language models the default starting point for almost every task. BART, Sentence-BERT, and Dense Passage Retrieval extended pretraining to generation, embeddings, and search. Prefix-Tuning, Self-Instruct, and TruthfulQA marked the shift to large language models (LLMs) and generative AI. The newest work, such as test-time scaling in s1, is about getting those models to reason. Keeping up is hard: ACL and EMNLP alone now publish thousands of papers a year. This reading list is a starting point: 100 of the most influential NLP and deep learning papers from 2016 to 2025, the work the research community has cited and built on most.

How the papers were selected

The papers were selected by citation count from the three flagship NLP and computational linguistics conferences: ACL, EMNLP, and NAACL. For each year from 2016 to 2025, we took the ten most-cited papers across that year’s conferences. Selecting by year keeps recent work from being crowded out by older papers that have simply had more time to collect citations. NAACL is not held every year, and there was no edition in 2017, 2020, or 2023, so those years draw from ACL and EMNLP only.

The list is ordered newest year first. Read from the bottom up, it doubles as a short history of how the field moved:

• neural machine translation, sequence labeling, and large reading-comprehension datasets
• pretrained language models
• prompting and parameter-efficient tuning
• instruction tuning and evaluating large language models
• today’s reasoning and agentic models

A note on scope: this list reflects our selection criteria, not a definitive list of the best NLP papers. There may well be better or more important papers elsewhere. Some landmark NLP work appeared outside these three conferences. The Transformer and GPT-3 were published at NeurIPS, T5 in JMLR, and many influential LLM technical reports exist only on arXiv. Citation counts also favor popular topics and lag behind the newest work. Treat the list as a well-grounded map of the field rather than the final word. For the full year-by-year lists, see the Most Influential ACL, EMNLP, and NAACL papers pages. Lists for other venues are on the Best Paper Digest page.

Keeping up beyond this list

If you work across fields, the companion lists follow the same method: 100 must-read computer vision papers of the past 10 years (CVPR, ICCV, ECCV) and 100 must-read machine learning papers of the past 10 years (NeurIPS, ICML, ICLR).

A citation-based list looks backward. It tells you what mattered, not what is emerging this month. Every paper on Paper Digest links to related papers, patents, grants, and experts, so you can explore the research around any of the work below. You can also run a literature review on a specific topic. If you would like new NLP and LLM papers matched to your interests each morning, you can sign up and set up a daily digest.

TABLE 1: Paper Digest: 100 Must-Read Natural Language Processing Papers of the Past 10 Years (2016-2025)