{"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/probe-pruning-accelerating-llms-through","title":"Probe Pruning: Accelerating LLMs through Dynamic Pruning via Model-Probing","arxiv_id":"2502.15618","date":"2025-02-21","proceeding":null,"authors":["Qi Le","Enmao Diao","Ziyan Wang","Xinran Wang","Jie Ding","Li Yang","Ali Anwar"],"abstract":"We introduce Probe Pruning (PP), a novel framework for online, dynamic, structured pruning of Large Language Models (LLMs) applied in a batch-wise manner. PP leverages the insight that not all samples and tokens contribute equally to the model's output, and probing a small portion of each batch effectively identifies crucial weights, enabling tailored dynamic pruning for different batches. It comprises three main stages: probing, history-informed pruning, and full inference. In the probing stage, PP selects a small yet crucial set of hidden states, based on residual importance, to run a few model layers ahead. During the history-informed pruning stage, PP strategically integrates the probing states with historical states. Subsequently, it structurally prunes weights based on the integrated states and the PP importance score, a metric developed specifically to assess the importance of each weight channel in maintaining performance. In the final stage, full inference is conducted on the remaining weights. A major advantage of PP is its compatibility with existing models, as it operates without requiring additional neural network modules or fine-tuning. Comprehensive evaluations of PP on LLaMA-2/3 and OPT models reveal that even minimal probing-using just 1.5% of FLOPs-can substantially enhance the efficiency of structured pruning of LLMs. For instance, when evaluated on LLaMA-2-7B with WikiText2, PP achieves a 2.56 times lower ratio of performance degradation per unit of runtime reduction compared to the state-of-the-art method at a 40% pruning ratio. Our code is available at https://github.com/Qi-Le1/Probe_Pruning.","url_abs":"https://arxiv.org/abs/2502.15618v1","url_pdf":"https://arxiv.org/pdf/2502.15618v1.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":"probe-pruning-accelerating-llms-through","repo_url":"https://github.com/qi-le1/probe_pruning","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}}],"tasks":[],"methods":[{"method_slug":"opt","method_name":"OPT"},{"method_slug":"pruning","method_name":"Pruning"},{"method_slug":"set","method_name":"SET"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/2502.15618","atlas_url":"https://app.syntology.ai/?focus=2502.15618","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2502.15618"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-25T09:33:49+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":"deterministic:regex_extraction","url":"https://github.com/Qi-Le1/Probe_Pruning","reach":{"status":"ok","spdx":"MIT"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/qi-le1/probe_pruning","reach":{"status":"ok","spdx":"MIT"}}],"summary":{"ran_fixture":1,"ran_draft_wrong":1,"unverified":1},"by_repo_kind":{"official":{"samples":2,"ran":1,"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":1,"samples":[{"code_sha256_prefix":"9b4dff79d5e6102c","entry":"apply_rotary_pos_emb","repo":"qi-le1/probe_pruning","repo_kind":"official","path":"src/LLMPruner/models/hf_llama/modeling_llama.py","file_url":"https://github.com/qi-le1/probe_pruning/blob/HEAD/src/LLMPruner/models/hf_llama/modeling_llama.py","link_basis":"first_harvest_node","language":"python","status":"ran_fixture","verification_level":1,"contract_check":"RAISES","metamorphic_tier":"invariant","behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"9b4dff79d5e6102c"}},{"code_sha256_prefix":"b99eea6376d1e212","entry":"rotate_half","repo":null,"repo_kind":null,"path":null,"file_url":null,"link_basis":"identical_code_first_harvested_elsewhere","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"MISDECLARED","metamorphic_tier":"invariant","behaviour_fingerprint":true,"licence":null,"inline_ok":false,"mcp_get_code":{"code_sha256":"b99eea6376d1e212"}},{"code_sha256_prefix":"30f55a1269244394","entry":"HiddenRepresentationPruning","repo":"qi-le1/probe_pruning","repo_kind":"official","path":"src/model/pruning_module.py","file_url":"https://github.com/qi-le1/probe_pruning/blob/HEAD/src/model/pruning_module.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"30f55a1269244394"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}