{"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/arxiv-2605-17932","title":"Prompt Compression in Diffusion Large Language Models: Evaluating LLMLingua-2 on LLaDA","arxiv_id":"2605.17932","date":"2026-05-18","proceeding":null,"authors":["Sterling Huang","Abigayle Brown","Jiyoo Noh","Jiakang Xu","Wantong Huo","Kaung Myat Kyaw","Jonathan Chan"],"abstract":"Prompt compression reduces inference cost and context length in large language models, but prior evaluations focus mainly on autoregressive architectures. This study examines whether LLMLingua-2 transfers effectively to diffusion large language models (DLLMs), specifically LLaDA-8B-Instruct. We evaluate GSM8K, DUC2004, and ShareGPT using 250 prompts per dataset at an approximate 50\\% compression ratio, covering mathematical reasoning, prompt reconstruction, and summarization. Outputs from original, compressed, and reconstructed prompts are compared using exact-match accuracy, BLEU, ROUGE, and BERTScore. Results show that high semantic preservation does not necessarily ensure stable downstream behavior in diffusion models. Summarization remains relatively robust, while mathematical reasoning degrades substantially despite high semantic similarity. Reconstruction further shows that semantically similar prompts may omit reasoning-critical information needed for stable denoising. Overall, compression failures are mainly driven by information omission rather than semantic drift, suggesting that autoregressive prompt compression methods may not transfer uniformly to DLLMs. These findings motivate diffusion-aware compression strategies.","url_abs":"https://arxiv.org/abs/2605.17932","url_pdf":"https://arxiv.org/pdf/2605.17932","source":{"archive":null,"snapshot":"2025-07-28","note":"not in the Papers with Code archive (frozen at the snapshot)","row_kind":"graph","title_abstract_authors_date":"arXiv metadata, CC0 1.0 (https://info.arxiv.org/help/license)"},"code_links":[],"tasks":[],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2605.17932","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2605.17932"}},"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":[{"mentioned_in_github":null,"is_official":null,"provenance":"deterministic:regex_extraction","mentioned_in_paper":null,"url":"https://github.com/axxx129-ops/DLLM-Compression","reach":null}],"summary":{"ran":9},"by_repo_kind":{"found_in_text":{"samples":9,"ran":9,"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":9,"samples":[{"code_sha256_prefix":"432d5314a4ebadbd","entry":"bleu_score","repo":"axxx129-ops/DLLM-Compression","repo_kind":"found_in_text","path":"evaluation/evaluate.py","file_url":"https://github.com/axxx129-ops/DLLM-Compression/blob/HEAD/evaluation/evaluate.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"432d5314a4ebadbd"}},{"code_sha256_prefix":"30e2ac18649d17d0","entry":"ensure_directory","repo":"axxx129-ops/DLLM-Compression","repo_kind":"found_in_text","path":"evaluation/gsm_evaluation.py","file_url":"https://github.com/axxx129-ops/DLLM-Compression/blob/HEAD/evaluation/gsm_evaluation.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"30e2ac18649d17d0"}},{"code_sha256_prefix":"435c82878848e0f5","entry":"extract_numeric","repo":"axxx129-ops/DLLM-Compression","repo_kind":"found_in_text","path":"evaluation/gsm_evaluation.py","file_url":"https://github.com/axxx129-ops/DLLM-Compression/blob/HEAD/evaluation/gsm_evaluation.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"435c82878848e0f5"}},{"code_sha256_prefix":"e1a1af066530cecd","entry":"fallback_bleu","repo":"axxx129-ops/DLLM-Compression","repo_kind":"found_in_text","path":"evaluation/evaluate.py","file_url":"https://github.com/axxx129-ops/DLLM-Compression/blob/HEAD/evaluation/evaluate.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"e1a1af066530cecd"}},{"code_sha256_prefix":"309e5d607624016b","entry":"load","repo":"axxx129-ops/DLLM-Compression","repo_kind":"found_in_text","path":"evaluation/check_exact_match.py","file_url":"https://github.com/axxx129-ops/DLLM-Compression/blob/HEAD/evaluation/check_exact_match.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"309e5d607624016b"}},{"code_sha256_prefix":"c67ec6398e695ef0","entry":"main","repo":"axxx129-ops/DLLM-Compression","repo_kind":"found_in_text","path":"evaluation/check_exact_match.py","file_url":"https://github.com/axxx129-ops/DLLM-Compression/blob/HEAD/evaluation/check_exact_match.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"c67ec6398e695ef0"}},{"code_sha256_prefix":"9e8868858c2b91b7","entry":"ngrams","repo":"axxx129-ops/DLLM-Compression","repo_kind":"found_in_text","path":"evaluation/evaluate.py","file_url":"https://github.com/axxx129-ops/DLLM-Compression/blob/HEAD/evaluation/evaluate.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"9e8868858c2b91b7"}},{"code_sha256_prefix":"6a07c38a08dc72e6","entry":"normalize","repo":"axxx129-ops/DLLM-Compression","repo_kind":"found_in_text","path":"evaluation/check_exact_match.py","file_url":"https://github.com/axxx129-ops/DLLM-Compression/blob/HEAD/evaluation/check_exact_match.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"6a07c38a08dc72e6"}},{"code_sha256_prefix":"8eb5d8d1d95ca3c0","entry":"numeric_equivalent","repo":"axxx129-ops/DLLM-Compression","repo_kind":"found_in_text","path":"evaluation/gsm_evaluation.py","file_url":"https://github.com/axxx129-ops/DLLM-Compression/blob/HEAD/evaluation/gsm_evaluation.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"8eb5d8d1d95ca3c0"}}]},"arxiv_metadata":{"licence":"arXiv metadata, CC0 1.0 (https://info.arxiv.org/help/license)","fields":["title","abstract","authors","date"],"primary_category":"cs.CL","source":"arxiv_2026.jsonl"},"syntology_extracted_results":null}