{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:O56Z7NAPGNCQCTHUE4Y64WKPDV","short_pith_number":"pith:O56Z7NAP","schema_version":"1.0","canonical_sha256":"777d9fb40f3345014cf42731ee594f1d6c3d4668474c6e7fc7c0c5195396e74e","source":{"kind":"arxiv","id":"2606.11854","version":1},"attestation_state":"computed","paper":{"title":"Fine-tuning Multi-modal LLMs with ART: Art-based Reinforcement Training","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Michal Chudoba, Petra Galuscakova, Sergey Alyaev, Tomasz Wiktorski","submitted_at":"2026-06-10T09:30:37Z","abstract_excerpt":"There are two main Parameter-Efficient Fine-Tuning (PEFT) techniques for Large Language Models (LLMs). While Low-Rank Adaptation (LoRA) introduces additional weights between the LLM layers, Soft Prompting introduces additional fine-tuning-specific raw tokens to an LLM input. However, both require modification to the computational graphs of precompiled, preoptimized LLMs. As a result, neither is fully supported in high-throughput engines like vLLM. We propose fine-tuning with ART (Art-based Reinforcement Training). The method injects information into a frozen Multimodal Large Language Model (ML"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2606.11854","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-06-10T09:30:37Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"83ad85e456f8780c89f234c6b3368d899196dedc8b5e059c52ef9107e9cb4865","abstract_canon_sha256":"16522dfbd7e7abd0b90849b375bd7e38aa2ff22bcc827b1341b2ca9180afe97c"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-11T01:10:11.885944Z","signature_b64":"thKzGN9uu5tfAs9KrGn1Tm5Tog8RB5YwG4bQl8+gY8qtQ9vnZ+m8aPvLKO8RiLFyZuzFSXSwOBMmIbP4pF2JCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"777d9fb40f3345014cf42731ee594f1d6c3d4668474c6e7fc7c0c5195396e74e","last_reissued_at":"2026-06-11T01:10:11.885437Z","signature_status":"signed_v1","first_computed_at":"2026-06-11T01:10:11.885437Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Fine-tuning Multi-modal LLMs with ART: Art-based Reinforcement Training","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Michal Chudoba, Petra Galuscakova, Sergey Alyaev, Tomasz Wiktorski","submitted_at":"2026-06-10T09:30:37Z","abstract_excerpt":"There are two main Parameter-Efficient Fine-Tuning (PEFT) techniques for Large Language Models (LLMs). While Low-Rank Adaptation (LoRA) introduces additional weights between the LLM layers, Soft Prompting introduces additional fine-tuning-specific raw tokens to an LLM input. However, both require modification to the computational graphs of precompiled, preoptimized LLMs. As a result, neither is fully supported in high-throughput engines like vLLM. We propose fine-tuning with ART (Art-based Reinforcement Training). The method injects information into a frozen Multimodal Large Language Model (ML"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.11854","kind":"arxiv","version":1},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2606.11854/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2606.11854","created_at":"2026-06-11T01:10:11.885509+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.11854v1","created_at":"2026-06-11T01:10:11.885509+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.11854","created_at":"2026-06-11T01:10:11.885509+00:00"},{"alias_kind":"pith_short_12","alias_value":"O56Z7NAPGNCQ","created_at":"2026-06-11T01:10:11.885509+00:00"},{"alias_kind":"pith_short_16","alias_value":"O56Z7NAPGNCQCTHU","created_at":"2026-06-11T01:10:11.885509+00:00"},{"alias_kind":"pith_short_8","alias_value":"O56Z7NAP","created_at":"2026-06-11T01:10:11.885509+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/O56Z7NAPGNCQCTHUE4Y64WKPDV","json":"https://pith.science/pith/O56Z7NAPGNCQCTHUE4Y64WKPDV.json","graph_json":"https://pith.science/api/pith-number/O56Z7NAPGNCQCTHUE4Y64WKPDV/graph.json","events_json":"https://pith.science/api/pith-number/O56Z7NAPGNCQCTHUE4Y64WKPDV/events.json","paper":"https://pith.science/paper/O56Z7NAP"},"agent_actions":{"view_html":"https://pith.science/pith/O56Z7NAPGNCQCTHUE4Y64WKPDV","download_json":"https://pith.science/pith/O56Z7NAPGNCQCTHUE4Y64WKPDV.json","view_paper":"https://pith.science/paper/O56Z7NAP","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.11854&json=true","fetch_graph":"https://pith.science/api/pith-number/O56Z7NAPGNCQCTHUE4Y64WKPDV/graph.json","fetch_events":"https://pith.science/api/pith-number/O56Z7NAPGNCQCTHUE4Y64WKPDV/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/O56Z7NAPGNCQCTHUE4Y64WKPDV/action/timestamp_anchor","attest_storage":"https://pith.science/pith/O56Z7NAPGNCQCTHUE4Y64WKPDV/action/storage_attestation","attest_author":"https://pith.science/pith/O56Z7NAPGNCQCTHUE4Y64WKPDV/action/author_attestation","sign_citation":"https://pith.science/pith/O56Z7NAPGNCQCTHUE4Y64WKPDV/action/citation_signature","submit_replication":"https://pith.science/pith/O56Z7NAPGNCQCTHUE4Y64WKPDV/action/replication_record"}},"created_at":"2026-06-11T01:10:11.885509+00:00","updated_at":"2026-06-11T01:10:11.885509+00:00"}