# Pairs

Source: [Pairs](<https://www.sota2.com/research/dataset/pairs>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

## Benchmarks

Page 1 of 6. This is an index page, not the full collection.

- [Next index page](<https://www.sota2.com/research/dataset/pairs.md?page=2>): Follow this link to discover more pages.

- [Paraphrase Identification on Quora Question Pairs (test)](<https://www.sota2.com/research/sota/paraphrase-identification-on-quora-question-pairs-test.md>): Paraphrase Identification; Quora Question Pairs (test).

- [Relative Pose Estimation on ScanNet 1500 pairs (test)](<https://www.sota2.com/research/sota/relative-pose-estimation-on-scannet-1500-pairs-test.md>): Relative Pose Estimation; ScanNet 1500 pairs (test).

- [Photorealistic Style Transfer on Novel content-style pairs](<https://www.sota2.com/research/sota/photorealistic-style-transfer-on-novel-content-style-pairs.md>): Photorealistic Style Transfer; Novel content-style pairs.

- [Image Matching on DeSurT (833 pairs total)](<https://www.sota2.com/research/sota/image-matching-on-desurt-833-pairs-total.md>): Image Matching; DeSurT (833 pairs total).

- [UI Generation on 429 human-rated UI pairs 1.0 (test)](<https://www.sota2.com/research/sota/ui-generation-on-429-human-rated-ui-pairs-1-0-test.md>): UI Generation; 429 human-rated UI pairs 1.0 (test).

- [Fairness Evaluation on CrowS-Pairs](<https://www.sota2.com/research/sota/fairness-evaluation-on-crows-pairs.md>): Fairness Evaluation; CrowS-Pairs.

- [Counterfactual Input Evaluation on CrowS-Pairs](<https://www.sota2.com/research/sota/counterfactual-input-evaluation-on-crows-pairs.md>): Counterfactual Input Evaluation; CrowS-Pairs.

- [Information Retrieval on MAIR 20% labeled query-document pairs](<https://www.sota2.com/research/sota/information-retrieval-on-mair-20-percent-labeled-query-document-pairs.md>): Information Retrieval; MAIR 20% labeled query-document pairs.

- [Domain Adaptation on Office+Caltech (DeCAF6 features, all pairs average)](<https://www.sota2.com/research/sota/domain-adaptation-on-office-plus-caltech-decaf6-features-all-pairs-average.md>): Domain Adaptation; Office+Caltech DeCAF6 features (various pairs).

- [Latency Measurement on OOL-Pairs for Long-context reasoning](<https://www.sota2.com/research/sota/latency-measurement-on-ool-pairs-for-long-context-reasoning.md>): Long-context reasoning; OOL-Pairs.

- [Long-context reasoning (Pairs) on OOL-Pairs](<https://www.sota2.com/research/sota/long-context-reasoning-pairs-on-ool-pairs.md>): Long-context reasoning (Pairs); OOL-Pairs.

- [Zero-shot Evaluation on ARC (c/e), BoolQ, CrowS-Pairs, OBQA, PIQA, RACE, SiQA, TruthfulQA, WinoGrande](<https://www.sota2.com/research/sota/zero-shot-evaluation-on-arc-c-e-boolq-crows-pairs-obqa-piqa-race-siqa-truthfulqa-winogrande.md>): Zero-shot Evaluation; ARC-Challenge, ARC-Easy, BoolQ, CrowS-Pairs, OpenBookQA, PIQA, RACE, SiQA, TruthfulQA, WinoGrande zero-shot.

- [Clinical Relatedness Detection on General Clinical Relation Pairs](<https://www.sota2.com/research/sota/clinical-relatedness-detection-on-general-clinical-relation-pairs.md>): Clinical Relatedness Detection; General Clinical Relation Pairs.

- [Clinical Similarity Detection on General Clinical Relation Pairs](<https://www.sota2.com/research/sota/clinical-similarity-detection-on-general-clinical-relation-pairs.md>): Clinical Similarity Detection; General Clinical Relation Pairs.

- [Multi-Session Place Recognition on SNAIL 2 Pairs](<https://www.sota2.com/research/sota/multi-session-place-recognition-on-snail-2-pairs.md>): Multi-Session Place Recognition; SNAIL 2 Pairs.

- [Multi-Session Place Recognition on HeLiPR 2 Pairs](<https://www.sota2.com/research/sota/multi-session-place-recognition-on-helipr-2-pairs.md>): Multi-Session Place Recognition; HeLiPR 2 Pairs.

- [Multi-Session Place Recognition on NCLT 3 Pairs](<https://www.sota2.com/research/sota/multi-session-place-recognition-on-nclt-3-pairs.md>): Multi-Session Place Recognition; NCLT 3 Pairs.

- [Sentence Interpolation Smoothness on ARGO randomly sampled 200 sentence pairs](<https://www.sota2.com/research/sota/sentence-interpolation-smoothness-on-argo-randomly-sampled-200-sentence-pairs.md>): Sentence Interpolation Smoothness; ARGO randomly sampled 200 sentence pairs.

- [Adaptive Retrieval Privacy Attack Success on 500 nearest-neighbor and random prefix pairs mid-network layer](<https://www.sota2.com/research/sota/adaptive-retrieval-privacy-attack-success-on-500-nearest-neighbor-and-random-prefix-pairs-mid-network-layer.md>): Adaptive Retrieval Privacy Attack Success; 500 nearest-neighbor and random prefix pairs mid-network layer.

- [LLM-as-a-Judge on High-contrast response pairs](<https://www.sota2.com/research/sota/llm-as-a-judge-on-high-contrast-response-pairs.md>): LLM-as-a-Judge; High-contrast response pairs.

- [Inter-session Place Recognition on Wild-Places Venman01-04 (12 query-database pairs)](<https://www.sota2.com/research/sota/inter-session-place-recognition-on-wild-places-venman01-04-12-query-database-pairs.md>): Inter-session place recognition; Wild-Places Venman01-04 (12 query-database pairs).

- [Dual-target drug design on 12,917 Dual-Target Pairs (test)](<https://www.sota2.com/research/sota/dual-target-drug-design-on-12-917-dual-target-pairs-test.md>): Dual-target drug design; 12,917 pairs of targets dual-target setting (test).

- [Adversarial Attack on 100 source-target pairs standard (test)](<https://www.sota2.com/research/sota/adversarial-attack-on-100-source-target-pairs-standard-test.md>): Adversarial Attack; 100 source-target pairs standard (test).

- [Religious Bias Evaluation on Multilingual CrowS-Pairs (test)](<https://www.sota2.com/research/sota/religious-bias-evaluation-on-multilingual-crows-pairs-test.md>): Religious Bias Evaluation; Multilingual CrowS-Pairs (test).

- [Racial Bias Evaluation on Multilingual CrowS-Pairs](<https://www.sota2.com/research/sota/racial-bias-evaluation-on-multilingual-crows-pairs.md>): Racial Bias Evaluation; Multilingual CrowS-Pairs racial bias.

- [Gender Bias Mitigation on Multilingual CrowS-Pairs](<https://www.sota2.com/research/sota/gender-bias-mitigation-on-multilingual-crows-pairs.md>): Gender Bias Mitigation; Multilingual CrowS-Pairs gender-sensitive attributes.

- [Relative Pose Estimation on MegaDepth 1500 outdoor pairs (test)](<https://www.sota2.com/research/sota/relative-pose-estimation-on-megadepth-1500-outdoor-pairs-test.md>): Relative Pose Estimation; MegaDepth 1500 outdoor pairs (test).

- [Relative Pose Estimation on MegaDepth 1500 pairs (test)](<https://www.sota2.com/research/sota/relative-pose-estimation-on-megadepth-1500-pairs-test.md>): Relative Pose Estimation; MegaDepth 1500 pairs (test).

- [Action Recognition on 3D Action Pairs](<https://www.sota2.com/research/sota/action-recognition-on-3d-action-pairs.md>): Action Recognition; 3D Action Pairs.

- [PMI ranking estimation on GoEmotions (500 held-out pairs)](<https://www.sota2.com/research/sota/pmi-ranking-estimation-on-goemotions-500-held-out-pairs.md>): PMI ranking estimation; GoEmotions (500 held-out pairs).

- [PMI ranking estimation on ChaosNLI (500 held-out pairs)](<https://www.sota2.com/research/sota/pmi-ranking-estimation-on-chaosnli-500-held-out-pairs.md>): PMI ranking estimation; ChaosNLI (500 held-out pairs).

- [PMI ranking estimation on Words (500 held-out pairs)](<https://www.sota2.com/research/sota/pmi-ranking-estimation-on-words-500-held-out-pairs.md>): PMI ranking estimation; Words (500 held-out pairs).

- [Targeted Adversarial Attack on ImageNet 1000 source-target pairs (val)](<https://www.sota2.com/research/sota/targeted-adversarial-attack-on-imagenet-1000-source-target-pairs-val.md>): Targeted Adversarial Attack; ImageNet 1000 source-target pairs (val).

- [Modeling Attack Resistance on PUF Challenge-Response Pairs](<https://www.sota2.com/research/sota/modeling-attack-resistance-on-puf-challenge-response-pairs.md>): Modeling Attack Resistance; PUF Challenge-Response Pairs.

- [Embodied Question Answering on BridgeEQA 1,100 QA pairs (test)](<https://www.sota2.com/research/sota/embodied-question-answering-on-bridgeeqa-1-100-qa-pairs-test.md>): Embodied Question Answering; BridgeEQA 1,100 QA pairs (test).

- [Antibody-Antigen Binding Affinity Ranking on 60k Subsampled Dataset (Random Split)](<https://www.sota2.com/research/sota/antibody-antigen-binding-affinity-ranking-on-60k-subsampled-dataset-random-split.md>): Antibody-Antigen Binding Affinity Ranking; Antibody-Antigen Binding Affinity Dataset 60,620 pairs subsampled (Random Split).

- [Metric Correlation with Human Judgments on Hearing-to-Translate five language pairs](<https://www.sota2.com/research/sota/metric-correlation-with-human-judgments-on-hearing-to-translate-five-language-pairs.md>): Metric Correlation with Human Judgments; Hearing-to-Translate five language pairs.

- [Retrieval on MAIR 10% labeled query-document pairs](<https://www.sota2.com/research/sota/retrieval-on-mair-10-percent-labeled-query-document-pairs.md>): Retrieval; MAIR 10% labeled query-document pairs.

- [Information Retrieval on MAIR 5% labeled query-document pairs (test)](<https://www.sota2.com/research/sota/information-retrieval-on-mair-5-percent-labeled-query-document-pairs-test.md>): Information Retrieval; MAIR 5% labeled query-document pairs (test).

- [Retrieval on MAIR 40% labeled query-document pairs](<https://www.sota2.com/research/sota/retrieval-on-mair-40-percent-labeled-query-document-pairs.md>): Retrieval; MAIR 40% labeled query-document pairs.

- [Question Answering on RepliQA 1,000 QA pairs (test)](<https://www.sota2.com/research/sota/question-answering-on-repliqa-1-000-qa-pairs-test.md>): Question Answering; RepliQA 1,000 QA pairs (test).

- [Text-to-Image Generation on MS-COCO 5k image-text pairs](<https://www.sota2.com/research/sota/text-to-image-generation-on-ms-coco-5k-image-text-pairs.md>): Text-to-Image; MS-COCO 5k image-text pairs.

- [Salient Object Detection on SIP 929 pairs (test)](<https://www.sota2.com/research/sota/salient-object-detection-on-sip-929-pairs-test.md>): Salient Object Detection; SIP 929 pairs (test).

- [Salient Object Detection on DUTLF-D 1,200 pairs (test)](<https://www.sota2.com/research/sota/salient-object-detection-on-dutlf-d-1-200-pairs-test.md>): Salient Object Detection; DUTLF-D 1,200 pairs (test).

- [Salient Object Detection on NLPR 1,000 pairs (test)](<https://www.sota2.com/research/sota/salient-object-detection-on-nlpr-1-000-pairs-test.md>): Salient Object Detection; NLPR 1,000 pairs (test).

- [Bias Evaluation on CrowS-Pairs (pct-stereotype)](<https://www.sota2.com/research/sota/bias-evaluation-on-crows-pairs-pct-stereotype.md>): Bias Evaluation; Crows-pairs.

- [MIL Explanation on Adjacent Pairs (test)](<https://www.sota2.com/research/sota/mil-explanation-on-adjacent-pairs-test.md>): MIL Explanation; Adjacent Pairs (test).

- [Image-to-Image Editing on ImagenHub Instruction-image pairs](<https://www.sota2.com/research/sota/image-to-image-editing-on-imagenhub-instruction-image-pairs.md>): Image-to-Image Editing; ImagenHub Instruction-image pairs.

- [Multiple Instance Learning classification on Adj. Pairs](<https://www.sota2.com/research/sota/multiple-instance-learning-classification-on-adj-pairs.md>): Multiple Instance Learning classification; Adj. Pairs.

- [Question Answering on Bamboogle 125 QA pairs](<https://www.sota2.com/research/sota/question-answering-on-bamboogle-125-qa-pairs.md>): Question Answering; Bamboogle 125 QA pairs.

- [Question Answering on HotpotQA 500 QA pairs](<https://www.sota2.com/research/sota/question-answering-on-hotpotqa-500-qa-pairs.md>): Question Answering; HotpotQA 500 QA pairs.

- [Question Answering on TriviaQA 500 QA pairs](<https://www.sota2.com/research/sota/question-answering-on-triviaqa-500-qa-pairs.md>): Question Answering; TriviaQA 500 QA pairs.

- [Infrared and Visible Image Fusion on IVOE (176 image pairs)](<https://www.sota2.com/research/sota/infrared-and-visible-image-fusion-on-ivoe-176-image-pairs.md>): Infrared and Visible Image Fusion; IVOE (176 image pairs).

- [Infrared and Visible Image Fusion on MSRS 361 image pairs (test)](<https://www.sota2.com/research/sota/infrared-and-visible-image-fusion-on-msrs-361-image-pairs-test.md>): Infrared and Visible Image Fusion; MSRS 361 image pairs (test).

- [Infrared and Visible Image Fusion on FMB 280 image pairs](<https://www.sota2.com/research/sota/infrared-and-visible-image-fusion-on-fmb-280-image-pairs.md>): Infrared and Visible Image Fusion; FMB 280 image pairs.

- [Cross-lingual sentence retrieval on Tatoeba Parallel 14 language pairs](<https://www.sota2.com/research/sota/cross-lingual-sentence-retrieval-on-tatoeba-parallel-14-language-pairs.md>): Cross-lingual sentence retrieval; Tatoeba Parallel 14 language pairs.

- [Paraphrase Identification on Quora Question Pairs](<https://www.sota2.com/research/sota/paraphrase-identification-on-quora-question-pairs.md>): Paraphrase Identification; Quora Question Pairs.

- [Paraphrase Identification on Quora Question Pairs (dev)](<https://www.sota2.com/research/sota/paraphrase-identification-on-quora-question-pairs-dev.md>): Paraphrase Identification; Quora Question Pairs (dev).

- [Overlap ratio estimation on Wiki (held-out pairs)](<https://www.sota2.com/research/sota/overlap-ratio-estimation-on-wiki-held-out-pairs.md>): Overlap ratio estimation; Wiki (held-out pairs).

- [Dynamic Feature Selection on SynPairs](<https://www.sota2.com/research/sota/dynamic-feature-selection-on-synpairs.md>): Dynamic Feature Selection; SynPairs.

- [Place Recognition on Wild-Places Karawatha01-04 (inter-session 12 query-database pairs)](<https://www.sota2.com/research/sota/place-recognition-on-wild-places-karawatha01-04-inter-session-12-query-database-pairs.md>): Place Recognition; Wild-Places Karawatha01-04 (inter-session (12 query-database pairs)).

- [Bias Evaluation on CrowS-Pairs](<https://www.sota2.com/research/sota/bias-evaluation-on-crows-pairs.md>): Bias Evaluation; CrowS-Pairs.

- [Overlap ratio estimation on Git (held-out pairs)](<https://www.sota2.com/research/sota/overlap-ratio-estimation-on-git-held-out-pairs.md>): Overlap ratio estimation; Git (held-out pairs).

- [Bivariate Causal Discovery on Tübingen cause–effect pairs](<https://www.sota2.com/research/sota/bivariate-causal-discovery-on-t-bingen-cause-effect-pairs.md>): Bivariate Causal Discovery; Tübingen cause–effect pairs.

- [Conditional Promoter DNA Sequence Design on Human Promoter Sequences Chromosomes 8 and 9 (test)](<https://www.sota2.com/research/sota/conditional-promoter-dna-sequence-design-on-human-promoter-sequences-chromosomes-8-and-9-test.md>): Conditional Promoter DNA Sequence Design; Human Promoter Sequences Chromosomes 8 and 9 100k sequences of 1024 base pairs (test).

- [Energy forecasting on Nine ordered (source, target) region pairs cross-region transfer](<https://www.sota2.com/research/sota/energy-forecasting-on-nine-ordered-source-target-region-pairs-cross-region-transfer.md>): Energy forecasting; Nine ordered (source, target) region pairs cross-region transfer.

- [Human Activity Recognition on USC-HAD source-target pairs](<https://www.sota2.com/research/sota/human-activity-recognition-on-usc-had-source-target-pairs.md>): Human Activity Recognition; USC-HAD source-target pairs.

- [Text-Guided High-Resolution Image Editing on 150 prompt-image pairs (test)](<https://www.sota2.com/research/sota/text-guided-high-resolution-image-editing-on-150-prompt-image-pairs-test.md>): Text-Guided High-Resolution Image Editing; 150 prompt-image pairs (test).

- [Spatial Question Answering on OST-Bench 1,396 QA pairs](<https://www.sota2.com/research/sota/spatial-question-answering-on-ost-bench-1-396-qa-pairs.md>): Spatial Question Answering; OST-Bench 1,396 QA pairs.

- [Value Alignment on Moral Integrity Corpus (MIC) (test)](<https://www.sota2.com/research/sota/value-alignment-on-moral-integrity-corpus-mic-test.md>): Value Alignment; Moral Integrity Corpus (MIC) 113.8k prompt-response pairs (test).

- [Text-to-Image Generation on Affordance Pairs Automatic Evaluation (test)](<https://www.sota2.com/research/sota/text-to-image-generation-on-affordance-pairs-automatic-evaluation-test.md>): Text-to-Image Generation; Affordance Pairs Automatic Evaluation (test).

- [Machine Translation on WMT 2022 Average (4 directions)](<https://www.sota2.com/research/sota/machine-translation-on-wmt-2022-average-4-directions.md>): Machine Translation; ALL Average of two language pairs in four directions wmt22-comet-da.

- [Rationale-label Pair Evaluation on Rationale-label pairs Pointwise evaluation](<https://www.sota2.com/research/sota/rationale-label-pair-evaluation-on-rationale-label-pairs-pointwise-evaluation.md>): Rationale-label Pair Evaluation; Rationale-label pairs Pointwise evaluation.

- [Binary Classification on STL-10 binary-pairs (test)](<https://www.sota2.com/research/sota/binary-classification-on-stl-10-binary-pairs-test.md>): Binary Classification; STL-10 binary-pairs (test).

- [Regular Language Recognition on Even Pairs](<https://www.sota2.com/research/sota/regular-language-recognition-on-even-pairs.md>): Regular Language Recognition; Even Pairs.

- [Information Extraction on ACL Anthology gold benchmark 100 query–chunk pairs](<https://www.sota2.com/research/sota/information-extraction-on-acl-anthology-gold-benchmark-100-query-chunk-pairs.md>): Information Extraction; ACL Anthology gold benchmark 100 query–chunk pairs.

- [LLM Judging on JudgeBench (GPT-4o 1.0 Generated Responses)](<https://www.sota2.com/research/sota/llm-judging-on-judgebench-gpt-4o-1-0-generated-responses.md>): LLM Judging; JudgeBench response pairs generated by GPT-4o 1.0.

- [Correlation with Human Judgment on 518 table pairs](<https://www.sota2.com/research/sota/correlation-with-human-judgment-on-518-table-pairs.md>): Correlation with Human Judgment; 518 table pairs.

- [Subject-to-video generation on OpenS2V-Nexus (held-out set of 180 subject-text pairs)](<https://www.sota2.com/research/sota/subject-to-video-generation-on-opens2v-nexus-held-out-set-of-180-subject-text-pairs.md>): Subject-to-video generation; OpenS2V-Nexus (held-out set of 180 subject-text pairs).

- [Object Replacement and Style Blending (800 pairs) (test)](<https://www.sota2.com/research/sota/object-replacement-and-style-blending-800-pairs-test.md>): Object Replacement and Style Blending; Object Replacement and Style Blending (800 pairs) (test).
