{"about":{"site":"https://codewithpapers.app","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.","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"},"url":"/paper/can-we-obtain-significant-success-in-rst","title":"Can we obtain significant success in RST discourse parsing by using Large Language Models?","arxiv_id":"2403.05065","date":"2024-03-08","proceeding":null,"authors":["Aru Maekawa","Tsutomu Hirao","Hidetaka Kamigaito","Manabu Okumura"],"abstract":"Recently, decoder-only pre-trained large language models (LLMs), with several tens of billion parameters, have significantly impacted a wide range of natural language processing (NLP) tasks. While encoder-only or encoder-decoder pre-trained language models have already proved to be effective in discourse parsing, the extent to which LLMs can perform this task remains an open research question. Therefore, this paper explores how beneficial such LLMs are for Rhetorical Structure Theory (RST) discourse parsing. Here, the parsing process for both fundamental top-down and bottom-up strategies is converted into prompts, which LLMs can work with. We employ Llama 2 and fine-tune it with QLoRA, which has fewer parameters that can be tuned. Experimental results on three benchmark datasets, RST-DT, Instr-DT, and the GUM corpus, demonstrate that Llama 2 with 70 billion parameters in the bottom-up strategy obtained state-of-the-art (SOTA) results with significant differences. Furthermore, our parsers demonstrated generalizability when evaluated on RST-DT, showing that, in spite of being trained with the GUM corpus, it obtained similar performances to those of existing parsers trained with RST-DT.","url_abs":"https://arxiv.org/abs/2403.05065v1","url_pdf":"https://arxiv.org/pdf/2403.05065v1.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":[{"paper_slug":"can-we-obtain-significant-success-in-rst","repo_url":"https://github.com/nttcslab-nlp/rstparser_eacl24","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"NOASSERTION"}}],"tasks":[{"task_slug":"decoder","task_name":"Decoder"},{"task_slug":"discourse-parsing","task_name":"Discourse Parsing"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/discourse-parsing-on-rst-dt","task":"Discourse Parsing","dataset":"RST-DT","model":"Bottom-up Llama 2 (70B)","rank_in_archive_order":1,"of":40,"metrics":{"Standard Parseval (Full)":"58.1","Standard Parseval (Nuclearity)":"70.4","Standard Parseval (Relation)":"60.0","Standard Parseval (Span)":"79.8"},"uses_additional_data":false},{"leaderboard":"/sota/discourse-parsing-on-rst-dt","task":"Discourse Parsing","dataset":"RST-DT","model":"Top-down Llama 2 (70B)","rank_in_archive_order":2,"of":40,"metrics":{"Standard Parseval (Full)":" 56.0","Standard Parseval (Nuclearity)":" 68.7","Standard Parseval (Relation)":"57.7","Standard Parseval (Span)":"78.8"},"uses_additional_data":false},{"leaderboard":"/sota/discourse-parsing-on-rst-dt","task":"Discourse Parsing","dataset":"RST-DT","model":"Bottom-up Llama 2 (13B)","rank_in_archive_order":3,"of":40,"metrics":{"Standard Parseval (Full)":"56.0","Standard Parseval (Nuclearity)":"68.1","Standard Parseval (Relation)":"57.8","Standard Parseval (Span)":" 78.3"},"uses_additional_data":false},{"leaderboard":"/sota/discourse-parsing-on-rst-dt","task":"Discourse Parsing","dataset":"RST-DT","model":"Bottom-up Llama 2 (7B)","rank_in_archive_order":4,"of":40,"metrics":{"Standard Parseval (Full)":"55.8","Standard Parseval (Nuclearity)":"67.5","Standard Parseval (Relation)":"57.6","Standard Parseval (Span)":" 78.2"},"uses_additional_data":false},{"leaderboard":"/sota/discourse-parsing-on-rst-dt","task":"Discourse Parsing","dataset":"RST-DT","model":"Top-down Llama 2 (13B)","rank_in_archive_order":6,"of":40,"metrics":{"Standard Parseval (Full)":" 55.6","Standard Parseval (Nuclearity)":"67.9","Standard Parseval (Relation)":"57.7","Standard Parseval (Span)":"78.6"},"uses_additional_data":false},{"leaderboard":"/sota/discourse-parsing-on-rst-dt","task":"Discourse Parsing","dataset":"RST-DT","model":"Top-down Llama 2 (7B)","rank_in_archive_order":13,"of":40,"metrics":{"Standard Parseval (Full)":"53.4","Standard Parseval (Nuclearity)":"65.4","Standard Parseval (Relation)":"55.2","Standard Parseval (Span)":"76.3"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2403.05065","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2403.05065"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-24T18:15:14+00:00","read_at_is":"when the build read Syntology's graph, not when any sample ran","claim":"Per-sample execution status on synthesized fixtures; not a correctness claim about the paper. Samples come from repositories linked to the paper, official or community; repo_kind says which.","repos":[{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/nttcslab-nlp/rstparser_eacl24","reach":{"status":"ok","spdx":"NOASSERTION"}}],"summary":{"unverified":2},"by_repo_kind":{"official":{"samples":2,"ran":0,"repositories":1}},"repo_kind_vocabulary":{"official":"The archive marks this repository official for the paper","named_in_paper":"The archive records that the paper mentions this repository; it is not marked official","listed":"In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper","found_in_text":"Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted","community":"Not in the archive's code links for this paper; a community repository Syntology harvested"},"n_pointer_only_for_licence":2,"samples":[{"code_sha256_prefix":"01eb601638ec987f","entry":"generate_answer","repo":"nttcslab-nlp/rstparser_eacl24","repo_kind":"official","path":"src/parse/parse_utils.py","file_url":"https://github.com/nttcslab-nlp/rstparser_eacl24/blob/HEAD/src/parse/parse_utils.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NOASSERTION","inline_ok":false,"mcp_get_code":{"code_sha256":"01eb601638ec987f"}},{"code_sha256_prefix":"f5d2ee40797d589c","entry":"split_span","repo":"nttcslab-nlp/rstparser_eacl24","repo_kind":"official","path":"src/parse/top_down/parse_top_down.py","file_url":"https://github.com/nttcslab-nlp/rstparser_eacl24/blob/HEAD/src/parse/top_down/parse_top_down.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NOASSERTION","inline_ok":false,"mcp_get_code":{"code_sha256":"f5d2ee40797d589c"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}