{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:RORN2USOLMRL6VS5HK7DDF6FD3","short_pith_number":"pith:RORN2USO","canonical_record":{"source":{"id":"2509.04549","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-09-04T17:32:56Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"b628015f45e4273f2ca0bddf9fafc6465dbf7a8f11c03e8939588825a89e089d","abstract_canon_sha256":"a0576881994695e20671ed43bb270c648e24e1811a18f0b925dd9f4754e648d3"},"schema_version":"1.0"},"canonical_sha256":"8ba2dd524e5b22bf565d3abe3197c51ec1882d352bf6113bf11e6af7fd5d4e50","source":{"kind":"arxiv","id":"2509.04549","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2509.04549","created_at":"2026-07-05T12:05:25Z"},{"alias_kind":"arxiv_version","alias_value":"2509.04549v1","created_at":"2026-07-05T12:05:25Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2509.04549","created_at":"2026-07-05T12:05:25Z"},{"alias_kind":"pith_short_12","alias_value":"RORN2USOLMRL","created_at":"2026-07-05T12:05:25Z"},{"alias_kind":"pith_short_16","alias_value":"RORN2USOLMRL6VS5","created_at":"2026-07-05T12:05:25Z"},{"alias_kind":"pith_short_8","alias_value":"RORN2USO","created_at":"2026-07-05T12:05:25Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:RORN2USOLMRL6VS5HK7DDF6FD3","target":"record","payload":{"canonical_record":{"source":{"id":"2509.04549","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-09-04T17:32:56Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"b628015f45e4273f2ca0bddf9fafc6465dbf7a8f11c03e8939588825a89e089d","abstract_canon_sha256":"a0576881994695e20671ed43bb270c648e24e1811a18f0b925dd9f4754e648d3"},"schema_version":"1.0"},"canonical_sha256":"8ba2dd524e5b22bf565d3abe3197c51ec1882d352bf6113bf11e6af7fd5d4e50","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T12:05:25.123016Z","signature_b64":"sBCZT0OsX80OACY0u5c+wwDCfV6PVNFMk73v106OFA/EkJTRUrhi4DJNPfo+JQjwOsCikACl2BOqG6elLlPIDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"8ba2dd524e5b22bf565d3abe3197c51ec1882d352bf6113bf11e6af7fd5d4e50","last_reissued_at":"2026-07-05T12:05:25.122597Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T12:05:25.122597Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2509.04549","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T12:05:25Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"BFwtYgYV96lmKTS383jE3k+yHJpb/McA3yoH+DZt9qpILzQ0I95lXrRUBhN0m6du1K2/mRirRoE7HHpcXleTCg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-22T01:36:37.175983Z"},"content_sha256":"2d9d8e852e53340ab0b7cc6cedd9f258f0a4fc3444c8d853b598e285d6bf1f82","schema_version":"1.0","event_id":"sha256:2d9d8e852e53340ab0b7cc6cedd9f258f0a4fc3444c8d853b598e285d6bf1f82"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:RORN2USOLMRL6VS5HK7DDF6FD3","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Manipulating Transformer-Based Models: Controllability, Steerability, and Robust Interventions","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Faruk Alpay, Taylan Alpay","submitted_at":"2025-09-04T17:32:56Z","abstract_excerpt":"Transformer-based language models excel in NLP tasks, but fine-grained control remains challenging. This paper explores methods for manipulating transformer models through principled interventions at three levels: prompts, activations, and weights. We formalize controllable text generation as an optimization problem addressable via prompt engineering, parameter-efficient fine-tuning, model editing, and reinforcement learning. We introduce a unified framework encompassing prompt-level steering, activation interventions, and weight-space edits. We analyze robustness and safety implications, incl"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2509.04549","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/2509.04549/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T12:05:25Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"iun0bwE2IKxGtkUE8oX7QFW/+fMQZppf1cO/K9ba0RWVk5+35ihaVwHH2audvy9UdEH1vBUbr3zkbB+/VmXVDg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-22T01:36:37.176362Z"},"content_sha256":"bddf4d65480b411eed68881a1945a686c2fd14aee655a6f4b32c5e2daff45fa8","schema_version":"1.0","event_id":"sha256:bddf4d65480b411eed68881a1945a686c2fd14aee655a6f4b32c5e2daff45fa8"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/RORN2USOLMRL6VS5HK7DDF6FD3/bundle.json","state_url":"https://pith.science/pith/RORN2USOLMRL6VS5HK7DDF6FD3/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/RORN2USOLMRL6VS5HK7DDF6FD3/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-08-22T01:36:37Z","links":{"resolver":"https://pith.science/pith/RORN2USOLMRL6VS5HK7DDF6FD3","bundle":"https://pith.science/pith/RORN2USOLMRL6VS5HK7DDF6FD3/bundle.json","state":"https://pith.science/pith/RORN2USOLMRL6VS5HK7DDF6FD3/state.json","well_known_bundle":"https://pith.science/.well-known/pith/RORN2USOLMRL6VS5HK7DDF6FD3/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:RORN2USOLMRL6VS5HK7DDF6FD3","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"a0576881994695e20671ed43bb270c648e24e1811a18f0b925dd9f4754e648d3","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-09-04T17:32:56Z","title_canon_sha256":"b628015f45e4273f2ca0bddf9fafc6465dbf7a8f11c03e8939588825a89e089d"},"schema_version":"1.0","source":{"id":"2509.04549","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2509.04549","created_at":"2026-07-05T12:05:25Z"},{"alias_kind":"arxiv_version","alias_value":"2509.04549v1","created_at":"2026-07-05T12:05:25Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2509.04549","created_at":"2026-07-05T12:05:25Z"},{"alias_kind":"pith_short_12","alias_value":"RORN2USOLMRL","created_at":"2026-07-05T12:05:25Z"},{"alias_kind":"pith_short_16","alias_value":"RORN2USOLMRL6VS5","created_at":"2026-07-05T12:05:25Z"},{"alias_kind":"pith_short_8","alias_value":"RORN2USO","created_at":"2026-07-05T12:05:25Z"}],"graph_snapshots":[{"event_id":"sha256:bddf4d65480b411eed68881a1945a686c2fd14aee655a6f4b32c5e2daff45fa8","target":"graph","created_at":"2026-07-05T12:05:25Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2509.04549/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Transformer-based language models excel in NLP tasks, but fine-grained control remains challenging. This paper explores methods for manipulating transformer models through principled interventions at three levels: prompts, activations, and weights. We formalize controllable text generation as an optimization problem addressable via prompt engineering, parameter-efficient fine-tuning, model editing, and reinforcement learning. We introduce a unified framework encompassing prompt-level steering, activation interventions, and weight-space edits. We analyze robustness and safety implications, incl","authors_text":"Faruk Alpay, Taylan Alpay","cross_cats":["cs.AI"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-09-04T17:32:56Z","title":"Manipulating Transformer-Based Models: Controllability, Steerability, and Robust Interventions"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2509.04549","kind":"arxiv","version":1},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:2d9d8e852e53340ab0b7cc6cedd9f258f0a4fc3444c8d853b598e285d6bf1f82","target":"record","created_at":"2026-07-05T12:05:25Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"a0576881994695e20671ed43bb270c648e24e1811a18f0b925dd9f4754e648d3","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-09-04T17:32:56Z","title_canon_sha256":"b628015f45e4273f2ca0bddf9fafc6465dbf7a8f11c03e8939588825a89e089d"},"schema_version":"1.0","source":{"id":"2509.04549","kind":"arxiv","version":1}},"canonical_sha256":"8ba2dd524e5b22bf565d3abe3197c51ec1882d352bf6113bf11e6af7fd5d4e50","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"8ba2dd524e5b22bf565d3abe3197c51ec1882d352bf6113bf11e6af7fd5d4e50","first_computed_at":"2026-07-05T12:05:25.122597Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T12:05:25.122597Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"sBCZT0OsX80OACY0u5c+wwDCfV6PVNFMk73v106OFA/EkJTRUrhi4DJNPfo+JQjwOsCikACl2BOqG6elLlPIDQ==","signature_status":"signed_v1","signed_at":"2026-07-05T12:05:25.123016Z","signed_message":"canonical_sha256_bytes"},"source_id":"2509.04549","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:2d9d8e852e53340ab0b7cc6cedd9f258f0a4fc3444c8d853b598e285d6bf1f82","sha256:bddf4d65480b411eed68881a1945a686c2fd14aee655a6f4b32c5e2daff45fa8"],"state_sha256":"4d38398a18fbd1687e081c21fc48b05aad1eb0926023a9d9336a6d8ee5a6c7f9"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"YWPtu62P0PmZjvfQYomJkTWPP7+hRc/oro/oNCLYQ5b4o+pmh+eVGDHVQu9Xo0dJnsCTGgHio+95M6GwmWjwAA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-22T01:36:37.178979Z","bundle_sha256":"22959c02614ec0937ee000dc6c6a4539d147d5b1a7ab8f1baf0ee47b5dc1faf0"}}