{"url":"/task/semantic-textual-similarity","name":"Semantic Textual Similarity","slug":"semantic-textual-similarity","description_markdown":"Semantic textual similarity deals with determining how similar two pieces of texts are.\r\nThis can take the form of assigning a score from 1 to 5. Related tasks are paraphrase or duplicate identification.\r\n\r\nImage source: [Learning Semantic Textual Similarity from Conversations](https://arxiv.org/pdf/1804.07754.pdf)","categories":[{"name":"Natural Language Processing","url":"/area/natural-language-processing"}],"source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","slug_source":"archive_url"},"counts":{"papers_tagged":2381,"papers_with_code":693,"benchmarks":13,"benchmark_tables_in_archive":13,"benchmark_tables_shown":13,"benchmark_tables_withheld_as_spam":0,"benchmark_definition":"a leaderboard table with at least one row; benchmark_tables_shown also counts the zero-row tables; benchmark_tables_in_archive adds the tables withheld as spam","datasets":19,"subtasks":2,"parent_tasks":0},"benchmarks":[{"leaderboard":"/sota/semantic-textual-similarity-on-sts-benchmark","slug":"semantic-textual-similarity-on-sts-benchmark","dataset":"STS Benchmark","dataset_url":"/dataset/sts-benchmark","rows_in_archive":66,"metrics":["Pearson Correlation","Spearman Correlation","Accuracy","Dev Pearson Correlation","Dev Spearman Correlation"],"first_row_in_archive_order":{"model":"MT-DNN-SMART","paper_title":"SMART: Robust and Efficient Fine-Tuning for Pre-trained Natural Language Models through Principled Regularized Optimization","paper_url":"/paper/smart-robust-and-efficient-fine-tuning-for","paper_date":"2019-11-08","arxiv_id":"1911.03437","code_links":[{"title":"namisan/mt-dnn","url":"https://github.com/namisan/mt-dnn"},{"title":"microsoft/MT-DNN","url":"https://github.com/microsoft/MT-DNN"},{"title":"archinetai/smart-pytorch","url":"https://github.com/archinetai/smart-pytorch"},{"title":"archinetai/vat-pytorch","url":"https://github.com/archinetai/vat-pytorch"},{"title":"cliang1453/camero","url":"https://github.com/cliang1453/camero"},{"title":"chunhuililili/mt_dnn","url":"https://github.com/chunhuililili/mt_dnn"}],"syntology":{"n":8,"n_ran":6,"n_unverified":2,"n_pointer_only":1}}},{"leaderboard":"/sota/semantic-textual-similarity-on-mrpc","slug":"semantic-textual-similarity-on-mrpc","dataset":"MRPC","dataset_url":"/dataset/mrpc","rows_in_archive":45,"metrics":["Accuracy","F1"],"first_row_in_archive_order":{"model":"MT-DNN-SMART","paper_title":"SMART: Robust and Efficient Fine-Tuning for Pre-trained Natural Language Models through Principled Regularized Optimization","paper_url":"/paper/smart-robust-and-efficient-fine-tuning-for","paper_date":"2019-11-08","arxiv_id":"1911.03437","code_links":[{"title":"namisan/mt-dnn","url":"https://github.com/namisan/mt-dnn"},{"title":"microsoft/MT-DNN","url":"https://github.com/microsoft/MT-DNN"},{"title":"archinetai/smart-pytorch","url":"https://github.com/archinetai/smart-pytorch"},{"title":"archinetai/vat-pytorch","url":"https://github.com/archinetai/vat-pytorch"},{"title":"cliang1453/camero","url":"https://github.com/cliang1453/camero"},{"title":"chunhuililili/mt_dnn","url":"https://github.com/chunhuililili/mt_dnn"}],"syntology":{"n":8,"n_ran":6,"n_unverified":2,"n_pointer_only":1}}},{"leaderboard":"/sota/semantic-textual-similarity-on-mteb","slug":"semantic-textual-similarity-on-mteb","dataset":"MTEB","dataset_url":"/dataset/mteb","rows_in_archive":30,"metrics":["Spearman Correlation"],"first_row_in_archive_order":{"model":"AnglE-UAE","paper_title":"AnglE-optimized Text Embeddings","paper_url":"/paper/angle-optimized-text-embeddings","paper_date":"2023-09-22","arxiv_id":"2309.12871","code_links":[{"title":"SeanLee97/AnglE","url":"https://github.com/SeanLee97/AnglE"},{"title":"4ai/bellm","url":"https://github.com/4ai/bellm"}],"syntology":null}},{"leaderboard":"/sota/semantic-textual-similarity-on-sick","slug":"semantic-textual-similarity-on-sick","dataset":"SICK","dataset_url":"/dataset/sick","rows_in_archive":22,"metrics":["Spearman Correlation"],"first_row_in_archive_order":{"model":"PromCSE-RoBERTa-large (0.355B)","paper_title":"Improved Universal Sentence Embeddings with Prompt-based Contrastive Learning and Energy-based Learning","paper_url":"/paper/deep-continuous-prompt-for-contrastive-1","paper_date":"2022-03-14","arxiv_id":"2203.06875","code_links":[{"title":"yjiangcm/promcse","url":"https://github.com/yjiangcm/promcse"}],"syntology":null}},{"leaderboard":"/sota/semantic-textual-similarity-on-sts13","slug":"semantic-textual-similarity-on-sts13","dataset":"STS13","dataset_url":"/dataset/sts-benchmark","rows_in_archive":22,"metrics":["Spearman Correlation"],"first_row_in_archive_order":{"model":"AnglE-LLaMA-7B","paper_title":"AnglE-optimized Text Embeddings","paper_url":"/paper/angle-optimized-text-embeddings","paper_date":"2023-09-22","arxiv_id":"2309.12871","code_links":[{"title":"SeanLee97/AnglE","url":"https://github.com/SeanLee97/AnglE"},{"title":"4ai/bellm","url":"https://github.com/4ai/bellm"}],"syntology":null}},{"leaderboard":"/sota/semantic-textual-similarity-on-sts14","slug":"semantic-textual-similarity-on-sts14","dataset":"STS14","dataset_url":"/dataset/sts-benchmark","rows_in_archive":21,"metrics":["Spearman Correlation"],"first_row_in_archive_order":{"model":"AnglE-LLaMA-13B","paper_title":"AnglE-optimized Text Embeddings","paper_url":"/paper/angle-optimized-text-embeddings","paper_date":"2023-09-22","arxiv_id":"2309.12871","code_links":[{"title":"SeanLee97/AnglE","url":"https://github.com/SeanLee97/AnglE"},{"title":"4ai/bellm","url":"https://github.com/4ai/bellm"}],"syntology":null}},{"leaderboard":"/sota/semantic-textual-similarity-on-sts12","slug":"semantic-textual-similarity-on-sts12","dataset":"STS12","dataset_url":"/dataset/sts-benchmark","rows_in_archive":20,"metrics":["Spearman Correlation"],"first_row_in_archive_order":{"model":"PromptEOL+CSE+OPT-13B","paper_title":"Scaling Sentence Embeddings with Large Language Models","paper_url":"/paper/scaling-sentence-embeddings-with-large","paper_date":"2023-07-31","arxiv_id":"2307.16645","code_links":[{"title":"kongds/scaling_sentemb","url":"https://github.com/kongds/scaling_sentemb"}],"syntology":{"n":4,"n_ran":4,"n_unverified":0,"n_pointer_only":4}}},{"leaderboard":"/sota/semantic-textual-similarity-on-sts15","slug":"semantic-textual-similarity-on-sts15","dataset":"STS15","dataset_url":"/dataset/sts-benchmark","rows_in_archive":20,"metrics":["Spearman Correlation"],"first_row_in_archive_order":{"model":"PromptEOL+CSE+LLaMA-30B","paper_title":"Scaling Sentence Embeddings with Large Language Models","paper_url":"/paper/scaling-sentence-embeddings-with-large","paper_date":"2023-07-31","arxiv_id":"2307.16645","code_links":[{"title":"kongds/scaling_sentemb","url":"https://github.com/kongds/scaling_sentemb"}],"syntology":{"n":4,"n_ran":4,"n_unverified":0,"n_pointer_only":4}}},{"leaderboard":"/sota/semantic-textual-similarity-on-sts16","slug":"semantic-textual-similarity-on-sts16","dataset":"STS16","dataset_url":"/dataset/sts-benchmark","rows_in_archive":20,"metrics":["Spearman Correlation"],"first_row_in_archive_order":{"model":"AnglE-LLaMA-7B-v2","paper_title":"AnglE-optimized Text Embeddings","paper_url":"/paper/angle-optimized-text-embeddings","paper_date":"2023-09-22","arxiv_id":"2309.12871","code_links":[{"title":"SeanLee97/AnglE","url":"https://github.com/SeanLee97/AnglE"},{"title":"4ai/bellm","url":"https://github.com/4ai/bellm"}],"syntology":null}},{"leaderboard":"/sota/semantic-textual-similarity-on-senteval","slug":"semantic-textual-similarity-on-senteval","dataset":"SentEval","dataset_url":"/dataset/senteval","rows_in_archive":6,"metrics":["MRPC","SICK-R","SICK-E","STS"],"first_row_in_archive_order":{"model":"GenSen","paper_title":"Learning General Purpose Distributed Sentence Representations via Large Scale Multi-task Learning","paper_url":"/paper/learning-general-purpose-distributed-sentence","paper_date":"2018-03-30","arxiv_id":"1804.00079","code_links":[{"title":"facebookresearch/InferSent","url":"https://github.com/facebookresearch/InferSent"},{"title":"facebookresearch/SentEval","url":"https://github.com/facebookresearch/SentEval"},{"title":"Maluuba/gensen","url":"https://github.com/Maluuba/gensen"},{"title":"najafmurtaza/Developing-Machine-Learning-Models-in-Flask","url":"https://github.com/najafmurtaza/Developing-Machine-Learning-Models-in-Flask"}],"syntology":{"n":3,"n_ran":3,"n_unverified":0,"n_pointer_only":3}}},{"leaderboard":"/sota/semantic-textual-similarity-on-cxc","slug":"semantic-textual-similarity-on-cxc","dataset":"CxC","dataset_url":"/dataset/cxc","rows_in_archive":4,"metrics":["avg ± std"],"first_row_in_archive_order":{"model":"PromCSE-RoBERTa-large (0.355B)","paper_title":"Improved Universal Sentence Embeddings with Prompt-based Contrastive Learning and Energy-based Learning","paper_url":"/paper/deep-continuous-prompt-for-contrastive-1","paper_date":"2022-03-14","arxiv_id":"2203.06875","code_links":[{"title":"yjiangcm/promcse","url":"https://github.com/yjiangcm/promcse"}],"syntology":null}},{"leaderboard":"/sota/semantic-textual-similarity-on-mrpc-dev","slug":"semantic-textual-similarity-on-mrpc-dev","dataset":"MRPC Dev","dataset_url":"/dataset/mrpc","rows_in_archive":2,"metrics":["Accuracy"],"first_row_in_archive_order":{"model":"Synthesizer (R+V)","paper_title":"Synthesizer: Rethinking Self-Attention in Transformer Models","paper_url":"/paper/synthesizer-rethinking-self-attention-in","paper_date":"2020-05-02","arxiv_id":"2005.00743","code_links":[{"title":"10-zin/Synthesizer","url":"https://github.com/10-zin/Synthesizer"}],"syntology":{"n":1,"n_ran":1,"n_unverified":0,"n_pointer_only":0}}},{"leaderboard":"/sota/semantic-textual-similarity-on-sick-r-1","slug":"semantic-textual-similarity-on-sick-r-1","dataset":"SICK-R","dataset_url":"/dataset/sick","rows_in_archive":2,"metrics":["Spearman Correlation"],"first_row_in_archive_order":{"model":"AnglE-LLaMA-7B","paper_title":"AnglE-optimized Text Embeddings","paper_url":"/paper/angle-optimized-text-embeddings","paper_date":"2023-09-22","arxiv_id":"2309.12871","code_links":[{"title":"SeanLee97/AnglE","url":"https://github.com/SeanLee97/AnglE"},{"title":"4ai/bellm","url":"https://github.com/4ai/bellm"}],"syntology":null}}],"datasets":[{"url":"/dataset/glue","name":"GLUE","full_name":"General Language Understanding Evaluation benchmark","num_papers_in_archive":3197},{"url":"/dataset/humaneval","name":"HumanEval","full_name":"","num_papers_in_archive":1201},{"url":"/dataset/mrpc","name":"MRPC","full_name":"Microsoft Research Paraphrase Corpus","num_papers_in_archive":786},{"url":"/dataset/sick","name":"SICK","full_name":"Sentences Involving Compositional Knowledge","num_papers_in_archive":348},{"url":"/dataset/senteval","name":"SentEval","full_name":"","num_papers_in_archive":168},{"url":"/dataset/mteb","name":"MTEB","full_name":"Massive Text Embedding Benchmark","num_papers_in_archive":155},{"url":"/dataset/emotion","name":"CARER","full_name":"Contextualized Affect Representations for Emotion Recognition","num_papers_in_archive":123},{"url":"/dataset/sts-benchmark","name":"STS Benchmark","full_name":"Semantic Textual Similarity","num_papers_in_archive":45},{"url":"/dataset/evalution","name":"EVALution","full_name":"EVALution","num_papers_in_archive":28},{"url":"/dataset/cxc","name":"CxC","full_name":"Crisscrossed Captions","num_papers_in_archive":22},{"url":"/dataset/pit","name":"PIT","full_name":"Paraphrase and Semantic Similarity in Twitter","num_papers_in_archive":22},{"url":"/dataset/multifc","name":"MultiFC","full_name":"","num_papers_in_archive":21},{"url":"/dataset/kornli","name":"KorNLI","full_name":null,"num_papers_in_archive":18},{"url":"/dataset/paranmt-50m","name":"PARANMT-50M","full_name":"","num_papers_in_archive":13},{"url":"/dataset/marc-ja","name":"JGLUE","full_name":"","num_papers_in_archive":7},{"url":"/dataset/sts-2014","name":"SemEval-2014 Task-10","full_name":"","num_papers_in_archive":6},{"url":"/dataset/gis","name":"GIS","full_name":"Github Issue Similarity","num_papers_in_archive":2},{"url":"/dataset/interpretable-sts","name":"Interpretable STS","full_name":"","num_papers_in_archive":1},{"url":"/dataset/czechnewsdatasetforsts","name":"CzechNewsDatasetForSTS","full_name":"","num_papers_in_archive":0}],"subtasks":[{"url":"/task/cross-lingual-semantic-textual-similarity","name":"Cross-Lingual Semantic Textual Similarity"},{"url":"/task/paraphrase-identification","name":"Paraphrase Identification"}],"parent_tasks":[],"papers":{"order":"repositories listed in the archive (desc), then date (desc); the archive holds no stars","population":"papers tagged with this task that list at least one repository in the archive","shown":30,"of":693,"tagged_in_all":2381,"items":[{"url":"/paper/bert-pre-training-of-deep-bidirectional","title":"BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding","date":"2018-10-11","arxiv_id":"1810.04805","repositories_listed":534,"syntology":{"n":659,"n_ran":204,"n_unverified":455,"n_pointer_only":149}},{"url":"/paper/roberta-a-robustly-optimized-bert-pretraining","title":"RoBERTa: A Robustly Optimized BERT Pretraining Approach","date":"2019-07-26","arxiv_id":"1907.11692","repositories_listed":67,"syntology":{"n":48,"n_ran":22,"n_unverified":26,"n_pointer_only":23}},{"url":"/paper/sentence-bert-sentence-embeddings-using","title":"Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks","date":"2019-08-27","arxiv_id":"1908.10084","repositories_listed":64,"syntology":{"n":58,"n_ran":20,"n_unverified":38,"n_pointer_only":11}},{"url":"/paper/exploring-the-limits-of-transfer-learning","title":"Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer","date":"2019-10-23","arxiv_id":"1910.10683","repositories_listed":57,"syntology":{"n":31,"n_ran":2,"n_unverified":29,"n_pointer_only":0}},{"url":"/paper/albert-a-lite-bert-for-self-supervised","title":"ALBERT: A Lite BERT for Self-supervised Learning of Language Representations","date":"2019-09-26","arxiv_id":"1909.11942","repositories_listed":48,"syntology":{"n":126,"n_ran":46,"n_unverified":80,"n_pointer_only":22}},{"url":"/paper/distilbert-a-distilled-version-of-bert","title":"DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter","date":"2019-10-02","arxiv_id":"1910.01108","repositories_listed":37,"syntology":{"n":27,"n_ran":19,"n_unverified":8,"n_pointer_only":0}},{"url":"/paper/xlnet-generalized-autoregressive-pretraining","title":"XLNet: Generalized Autoregressive Pretraining for Language Understanding","date":"2019-06-19","arxiv_id":"1906.08237","repositories_listed":27,"syntology":{"n":24,"n_ran":10,"n_unverified":14,"n_pointer_only":3}},{"url":"/paper/universal-sentence-encoder","title":"Universal Sentence Encoder","date":"2018-03-29","arxiv_id":"1803.11175","repositories_listed":24,"syntology":{"n":22,"n_ran":1,"n_unverified":21,"n_pointer_only":1}},{"url":"/paper/simcse-simple-contrastive-learning-of","title":"SimCSE: Simple Contrastive Learning of Sentence Embeddings","date":"2021-04-18","arxiv_id":"2104.08821","repositories_listed":23,"syntology":{"n":30,"n_ran":17,"n_unverified":13,"n_pointer_only":19}},{"url":"/paper/supervised-learning-of-universal-sentence","title":"Supervised Learning of Universal Sentence Representations from Natural Language Inference Data","date":"2017-05-05","arxiv_id":"1705.02364","repositories_listed":23,"syntology":{"n":7,"n_ran":6,"n_unverified":1,"n_pointer_only":7}},{"url":"/paper/ernie-enhanced-representation-through","title":"ERNIE: Enhanced Representation through Knowledge Integration","date":"2019-04-19","arxiv_id":"1904.09223","repositories_listed":19,"syntology":{"n":7,"n_ran":0,"n_unverified":7,"n_pointer_only":0}},{"url":"/paper/big-bird-transformers-for-longer-sequences","title":"Big Bird: Transformers for Longer Sequences","date":"2020-07-28","arxiv_id":"2007.14062","repositories_listed":14,"syntology":{"n":15,"n_ran":10,"n_unverified":5,"n_pointer_only":11}},{"url":"/paper/deberta-decoding-enhanced-bert-with","title":"DeBERTa: Decoding-enhanced BERT with Disentangled Attention","date":"2020-06-05","arxiv_id":"2006.03654","repositories_listed":14,"syntology":{"n":13,"n_ran":4,"n_unverified":9,"n_pointer_only":3}},{"url":"/paper/improving-language-understanding-by","title":"Improving Language Understanding by Generative Pre-Training","date":"2018-06-11","arxiv_id":null,"repositories_listed":13,"syntology":null},{"url":"/paper/fnet-mixing-tokens-with-fourier-transforms","title":"FNet: Mixing Tokens with Fourier Transforms","date":"2021-05-09","arxiv_id":"2105.03824","repositories_listed":12,"syntology":{"n":2,"n_ran":2,"n_unverified":0,"n_pointer_only":1}},{"url":"/paper/nystromformer-a-nystrom-based-algorithm-for","title":"Nyströmformer: A Nyström-Based Algorithm for Approximating Self-Attention","date":"2021-02-07","arxiv_id":"2102.03902","repositories_listed":10,"syntology":{"n":2,"n_ran":1,"n_unverified":1,"n_pointer_only":1}},{"url":"/paper/190910351","title":"TinyBERT: Distilling BERT for Natural Language Understanding","date":"2019-09-23","arxiv_id":"1909.10351","repositories_listed":10,"syntology":{"n":4,"n_ran":0,"n_unverified":4,"n_pointer_only":4}},{"url":"/paper/mobilebert-a-compact-task-agnostic-bert-for","title":"MobileBERT: a Compact Task-Agnostic BERT for Resource-Limited Devices","date":"2020-04-06","arxiv_id":"2004.02984","repositories_listed":7,"syntology":null},{"url":"/paper/tsdae-using-transformer-based-sequential","title":"TSDAE: Using Transformer-based Sequential Denoising Auto-Encoder for Unsupervised Sentence Embedding Learning","date":"2021-04-14","arxiv_id":"2104.06979","repositories_listed":6,"syntology":{"n":4,"n_ran":0,"n_unverified":4,"n_pointer_only":0}},{"url":"/paper/language-agnostic-bert-sentence-embedding","title":"Language-agnostic BERT Sentence Embedding","date":"2020-07-03","arxiv_id":"2007.01852","repositories_listed":6,"syntology":null},{"url":"/paper/squeezebert-what-can-computer-vision-teach","title":"SqueezeBERT: What can computer vision teach NLP about efficient neural networks?","date":"2020-06-19","arxiv_id":"2006.11316","repositories_listed":6,"syntology":{"n":1,"n_ran":0,"n_unverified":1,"n_pointer_only":0}},{"url":"/paper/smart-robust-and-efficient-fine-tuning-for","title":"SMART: Robust and Efficient Fine-Tuning for Pre-trained Natural Language Models through Principled Regularized Optimization","date":"2019-11-08","arxiv_id":"1911.03437","repositories_listed":6,"syntology":{"n":8,"n_ran":6,"n_unverified":2,"n_pointer_only":1}},{"url":"/paper/spanbert-improving-pre-training-by","title":"SpanBERT: Improving Pre-training by Representing and Predicting Spans","date":"2019-07-24","arxiv_id":"1907.10529","repositories_listed":6,"syntology":{"n":15,"n_ran":3,"n_unverified":12,"n_pointer_only":4}},{"url":"/paper/mteb-massive-text-embedding-benchmark","title":"MTEB: Massive Text Embedding Benchmark","date":"2022-10-13","arxiv_id":"2210.07316","repositories_listed":5,"syntology":{"n":13,"n_ran":3,"n_unverified":10,"n_pointer_only":0}},{"url":"/paper/informer-transformer-likes-informed-attention","title":"RealFormer: Transformer Likes Residual Attention","date":"2020-12-21","arxiv_id":"2012.11747","repositories_listed":5,"syntology":null},{"url":"/paper/q8bert-quantized-8bit-bert","title":"Q8BERT: Quantized 8Bit BERT","date":"2019-10-14","arxiv_id":"1910.06188","repositories_listed":5,"syntology":{"n":11,"n_ran":3,"n_unverified":8,"n_pointer_only":3}},{"url":"/paper/medsts-a-resource-for-clinical-semantic","title":"MedSTS: A Resource for Clinical Semantic Textual Similarity","date":"2018-08-28","arxiv_id":"1808.09397","repositories_listed":5,"syntology":null},{"url":"/paper/one-embedder-any-task-instruction-finetuned","title":"One Embedder, Any Task: Instruction-Finetuned Text Embeddings","date":"2022-12-19","arxiv_id":"2212.09741","repositories_listed":4,"syntology":{"n":20,"n_ran":10,"n_unverified":10,"n_pointer_only":0}},{"url":"/paper/llm-int8-8-bit-matrix-multiplication-for","title":"LLM.int8(): 8-bit Matrix Multiplication for Transformers at Scale","date":"2022-08-15","arxiv_id":"2208.07339","repositories_listed":4,"syntology":{"n":5,"n_ran":2,"n_unverified":3,"n_pointer_only":0}},{"url":"/paper/klue-korean-language-understanding-evaluation","title":"KLUE: Korean Language Understanding Evaluation","date":"2021-05-20","arxiv_id":"2105.09680","repositories_listed":4,"syntology":null}],"syntology_records":24,"syntology_note":"a paper without a record is not a recorded non-run: it may lack an arXiv id or simply be absent from the graph layer"},"description_links":{"kept":0,"unwrapped_to_text":0,"bare_urls_linked":0,"relative_images_dropped":0,"rule":"internal links are kept only when the target slug exists in the catalog"},"syntology":{"read_at":"2026-09-24T18:15:14+00:00","claim":"Per-sample execution status on synthesized fixtures ('ran N of M samples'); not a correctness claim and not a ranking signal.","status_vocabulary":{"ran_honours":"ran, honoured the contract we drafted","ran_violates":"ran, violated the contract we drafted","ran_draft_wrong":"ran; our contract draft was wrong, not the code","ran_fixture":"ran; our fixture could not drive it","ran":"ran on a synthesized input","unverified":"unverified (harvested, no recorded run)"}},"not_shown":{"libraries":"the archive has no per-task library table","trend_sparklines":"the Trend column of the benchmarks table was a rendered image; it is not in the archive","social_and_latest_sorts":"stars and social signals are not in the archive"}}