{"about":{"non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","site":"https://codewithpapers.app","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page","syntology":{"site":"https://syntology.ai","developers":"https://syntology.ai/developers","mcp":{"server":"https://syntology.ai/mcp","transport":"streamable-http","server_card":"https://syntology.ai/.well-known/mcp/server-card.json","auth":{"type":"trial token, no account","trial_token":"https://syntology.ai/api/oauth/trial/token","method":"POST","docs":"https://syntology.ai/developers"}},"have":"https://syntology.ai/api/graph/have?x=<method, arXiv id or title> (free, answers coverage only)","paper_base":"https://syntology.ai/paper/","atlas_base":"https://app.syntology.ai/?focus="},"machine_readable":[{"url":"https://codewithpapers.app/llms.txt","what":"the machine catalog: every machine-readable file, counted"},{"url":"https://codewithpapers.app/index/manifest.json","what":"paper-to-code index by arXiv id, with Syntology's counts"},{"url":"https://codewithpapers.app/search/manifest.json","what":"site search index (titles, authors) and its files"},{"url":"https://codewithpapers.app/download","what":"bulk files: Syntology's layer, described there"},{"url":"https://codewithpapers.app/build_manifest.json","what":"the build record: inputs, counts, exclusions, probes"}]},"url":"/paper/htlm-hyper-text-pre-training-and-prompting-of","title":"HTLM: Hyper-Text Pre-Training and Prompting of Language Models","arxiv_id":"2107.06955","date":"2021-07-14","proceeding":"ICLR 2022 4","authors":["Armen Aghajanyan","Dmytro Okhonko","Mike Lewis","Mandar Joshi","Hu Xu","Gargi Ghosh","Luke Zettlemoyer"],"abstract":"We introduce HTLM, a hyper-text language model trained on a large-scale web crawl. Modeling hyper-text has a number of advantages: (1) it is easily gathered at scale, (2) it provides rich document-level and end-task-adjacent supervision (e.g. class and id attributes often encode document category information), and (3) it allows for new structured prompting that follows the established semantics of HTML (e.g. to do zero-shot summarization by infilling title tags for a webpage that contains the input text). We show that pretraining with a BART-style denoising loss directly on simplified HTML provides highly effective transfer for a wide range of end tasks and supervision levels. HTLM matches or exceeds the performance of comparably sized text-only LMs for zero-shot prompting and fine-tuning for classification benchmarks, while also setting new state-of-the-art performance levels for zero-shot summarization. We also find that hyper-text prompts provide more value to HTLM, in terms of data efficiency, than plain text prompts do for existing LMs, and that HTLM is highly effective at auto-prompting itself, by simply generating the most likely hyper-text formatting for any available training data. We will release all code and models to support future HTLM research.","url_abs":"https://arxiv.org/abs/2107.06955v1","url_pdf":"https://arxiv.org/pdf/2107.06955v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[],"tasks":[{"task_slug":"data-to-text-generation","task_name":"Data-to-Text Generation"},{"task_slug":"denoising","task_name":"Denoising"},{"task_slug":"language-modelling","task_name":"Language Modelling"},{"task_slug":"table-to-text-generation","task_name":"Table-to-Text Generation"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/data-to-text-generation-on-webnlg","task":"Data-to-Text Generation","dataset":"WebNLG","model":"HTML (fine-tuning)","rank_in_archive_order":8,"of":20,"metrics":{"BLEU":"65.4"},"uses_additional_data":true},{"leaderboard":"/sota/data-to-text-generation-on-webnlg-full-1","task":"Data-to-Text Generation","dataset":"WebNLG Full","model":"HTLM (prefix 0.1%)","rank_in_archive_order":6,"of":8,"metrics":{"BLEU":"56.3"},"uses_additional_data":true},{"leaderboard":"/sota/table-to-text-generation-on-dart","task":"Table-to-Text Generation","dataset":"DART","model":"HTLM (fine-tuning)","rank_in_archive_order":5,"of":6,"metrics":{"BERT":"0.94","BLEU":"47.2","BLEURT":"0.4","METEOR":"0.39","Mover":"0.51","TER":"0.44"},"uses_additional_data":false},{"leaderboard":"/sota/table-to-text-generation-on-dart","task":"Table-to-Text Generation","dataset":"DART","model":"GPT-2-Large (fine-tuning)","rank_in_archive_order":6,"of":6,"metrics":{"BERT":"0.94","BLEU":"47.0","BLEURT":"0.4","METEOR":"0.39","Mover":"0.51","TER":"0.46"},"uses_additional_data":false},{"leaderboard":"/sota/table-to-text-generation-on-e2e","task":"Table-to-Text Generation","dataset":"E2E","model":"HTLM (fine-tuning)","rank_in_archive_order":1,"of":2,"metrics":{"BLEU":"70.3","CIDEr":"2.47","METEOR":"46.3","NIST":"8.90","ROUGE-L":"70.8"},"uses_additional_data":false},{"leaderboard":"/sota/table-to-text-generation-on-e2e","task":"Table-to-Text Generation","dataset":"E2E","model":"GPT-2-Large (fine-tuning)","rank_in_archive_order":2,"of":2,"metrics":{"BLEU":"68.5","CIDEr":"2.45","METEOR":"46.0","NIST":"8.78","ROUGE-L":"69.9"},"uses_additional_data":false},{"leaderboard":"/sota/table-to-text-generation-on-webnlg-all","task":"Table-to-Text Generation","dataset":"WebNLG (All)","model":"HTLM (fine-tuning)","rank_in_archive_order":1,"of":2,"metrics":{"BLEU":"55.6","METEOR":"0.42","TER":"0.4"},"uses_additional_data":false},{"leaderboard":"/sota/table-to-text-generation-on-webnlg-all","task":"Table-to-Text Generation","dataset":"WebNLG (All)","model":"GPT-2-Large (fine-tuning)","rank_in_archive_order":2,"of":2,"metrics":{"BLEU":"55.5","METEOR":"0.42","TER":"0.42"},"uses_additional_data":false},{"leaderboard":"/sota/table-to-text-generation-on-webnlg-seen","task":"Table-to-Text Generation","dataset":"WebNLG (Seen)","model":"HTLM (fine-tuning)","rank_in_archive_order":1,"of":2,"metrics":{"BLEU":"65.4","METEOR":"0.46","TER":"0.33"},"uses_additional_data":false},{"leaderboard":"/sota/table-to-text-generation-on-webnlg-seen","task":"Table-to-Text Generation","dataset":"WebNLG (Seen)","model":"GPT-2-Large (fine-tuning)","rank_in_archive_order":2,"of":2,"metrics":{"BLEU":"65.3","METEOR":"0.46","TER":"0.33"},"uses_additional_data":false},{"leaderboard":"/sota/table-to-text-generation-on-webnlg-unseen","task":"Table-to-Text Generation","dataset":"WebNLG (Unseen)","model":"HTLM (fine-tuning)","rank_in_archive_order":1,"of":2,"metrics":{"BLEU":"48.4","METEOR":"0.39","TER":"0.51"},"uses_additional_data":false},{"leaderboard":"/sota/table-to-text-generation-on-webnlg-unseen","task":"Table-to-Text Generation","dataset":"WebNLG (Unseen)","model":"GPT-2-Large (fine-tuning)","rank_in_archive_order":2,"of":2,"metrics":{"BLEU":"43.1","METEOR":"0.38","TER":"0.53"},"uses_additional_data":false}],"syntology":{"syntology_url":"https://syntology.ai/paper/2107.06955","atlas_url":"https://app.syntology.ai/?focus=2107.06955","mcp":null,"developers":"https://syntology.ai/developers","held":true,"reason":"arxiv_id_in_atlas","checked_against":{"atlas_ids":549449,"atlas_ids_sha256":"afd72399dafd9906e9062927f70655952385a3d75488558e1f0484e950ab09b7","read_at":"2026-10-05T01:46:21Z"},"agent_calls":[{"tool":"get_citation_path","arguments":{"paper_1":"2107.06955"},"arguments_in":{"paper_2":"another paper's arXiv id or title"},"call":"get_citation_path(paper_1=\"2107.06955\", paper_2=\"…\")"},{"tool":"get_concepts_for_paper","arguments":{"arxiv_id":"2107.06955"},"call":"get_concepts_for_paper(arxiv_id=\"2107.06955\")"}],"by_repo":[],"by_repo_is":"one row per repository in the archive's code links, Syntology's graph links or the samples; samples 0 means no sample is linked to this paper, a statement about Syntology's coverage, not about the repository: with harvested_for_other_papers present, Syntology harvested the repository for that many other papers, and without it, Syntology harvested nothing from it; placed_by_identical_code counts samples whose link records no repository, placed here because identical code was harvested from this repository for another paper; ran counts executions on a synthesized input and instrument counts failures of Syntology's instrument, not of the code"},"code_links_note":"the archive holds no code link for this paper","arxiv_metadata":null,"syntology_extracted_results":null,"record_sha256":"a70a7f9b37d885ccd552d2ef3008d15ac276ede1f6c6ccaefe7b2ad33d86c594","record_changed_at":"2026-09-28","record_changed_at_basis":"first_hashed"}