AI Paper Atlas

Research map / Benchmarks / SQuAD 2.0

SQuAD 2.0 leaderboard

22 papers at ICML, NeurIPS, ICLR, CVPR, AAAI, ICCV and ECCV, and the journals TPAMI, IJCV and JMLR report results on SQuAD 2.0. Best F1 (dev): 91.2 (GloSA-sum, ICLR 2026).

Numbers are as each paper reports them, read by AI from its results tables and not checked by hand. Papers can differ in setting (training data, model size, extra data); the topic column shows what each paper is about. Open the paper before you compare or quote a number.

SQuAD 2.0 · F1 (dev)

13 papers, each with its best reported number for its own method; higher is better.

3 more papers on this leaderboard, the baselines reported on it and a CSV download are free with a scime account.
Sign in free to see all

SQuAD 2.0 · F1

12 papers, each with its best reported number for its own method; higher is better.

2 more papers on this leaderboard, the baselines reported on it and a CSV download are free with a scime account.
Sign in free to see all

SQuAD 2.0 · EM (dev)

11 papers, each with its best reported number for its own method; higher is better.

#EM ↑MethodPaperTopic
188ELECTRA-1.75MELECTRA: Pre-training Text Encoders as Discriminators Rather Than Generators
ICLR 2020 · SQuAD results table
Attention and transformers
288SG-Net on ALBERT + Verifier
ALBERT
SG-Net: Syntax Guided Transformer for Language Representation
TPAMI 2020 · Table 1
Attention and transformers
387.9XLNetXLNet: Generalized Autoregressive Pretraining for Language Understanding
NeurIPS 2019 · Table 3
Language, safety and interpretability
487.8Retro-Reader (RV) on ALBERT
ALBERT
Retrospective Reader for Machine Reading Comprehension
AAAI 2021 · Table 3
Retrieval and knowledge
587.75ITC Joint
ALBERT
Text Compression-aided Transformer Encoding
TPAMI 2021 · Table 6
Attention and transformers
687.6B10-10-10
ELECTRA Large
Funnel-Transformer: Filtering out Sequential Redundancy for Efficient Language Processing
NeurIPS 2020 · Table 4
Attention and transformers
785.1SG-Net
BERT whole word masking
SG-Net: Syntax-Guided Machine Reading Comprehension
AAAI 2020 · Table (SQuAD 2.0 results)
Attention and transformers
882.8SAIN BERT large
BERT large
Semantics-Aware Inferential Network for Natural Language Understanding
AAAI 2021 · Table 3
Language, safety and interpretability
982ExLM
BERT base-sized
ExLM: Rethinking the Impact of $\texttt{[MASK]}$ Tokens in Masked Language Models
ICML 2025 · GLUE and SQuAD 2.0 results tab
Language, safety and interpretability
1080.9SemBERT_LARGE
BERT_LARGE
Semantics-Aware BERT for Language Understanding
AAAI 2020 · SNLI and SQuAD analysis table
Language, safety and interpretability
1 more papers on this leaderboard, the baselines reported on it and a CSV download are free with a scime account.
Sign in free to see all

SQuAD 2.0 · EM

10 papers, each with its best reported number for its own method; higher is better.

Not affiliated with any conference, journal or the benchmark's authors. Corrections: contact@latenticlabs.com.