{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:7SNFBXEIXODAJYKXKQM2UTNL3H","short_pith_number":"pith:7SNFBXEI","schema_version":"1.0","canonical_sha256":"fc9a50dc88bb8604e1575419aa4dabd9cb0a40fb9a22c44eeb3d5ec8f5819352","source":{"kind":"arxiv","id":"2401.12070","version":3},"attestation_state":"computed","paper":{"title":"Spotting LLMs With Binoculars: Zero-Shot Detection of Machine-Generated Text","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Abhimanyu Hans, Aniruddha Saha, Avi Schwarzschild, Hamid Kazemi, Jonas Geiping, Micah Goldblum, Tom Goldstein, Valeriia Cherepanova","submitted_at":"2024-01-22T16:09:47Z","abstract_excerpt":"Detecting text generated by modern large language models is thought to be hard, as both LLMs and humans can exhibit a wide range of complex behaviors. However, we find that a score based on contrasting two closely related language models is highly accurate at separating human-generated and machine-generated text. Based on this mechanism, we propose a novel LLM detector that only requires simple calculations using a pair of pre-trained LLMs. The method, called Binoculars, achieves state-of-the-art accuracy without any training data. It is capable of spotting machine text from a range of modern "},"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":"2401.12070","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-01-22T16:09:47Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"fd8652e9396076fef412d58b7febaac05a4f8d366e9a683b5e6c67b9f45194c0","abstract_canon_sha256":"d552c7d2fa5b027ec26981d3d687b1e1dd619953995f77315913ff406b8d34e8"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:19:53.650115Z","signature_b64":"wugQLuHD06wM6dJHe7Pg47/MG7et08MtMXizL5LRz8E87VWngO3YZnCjFsE/xkMNrRTXVDuwRMMWuLSRJr2nCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"fc9a50dc88bb8604e1575419aa4dabd9cb0a40fb9a22c44eeb3d5ec8f5819352","last_reissued_at":"2026-07-05T09:19:53.649621Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:19:53.649621Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Spotting LLMs With Binoculars: Zero-Shot Detection of Machine-Generated Text","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Abhimanyu Hans, Aniruddha Saha, Avi Schwarzschild, Hamid Kazemi, Jonas Geiping, Micah Goldblum, Tom Goldstein, Valeriia Cherepanova","submitted_at":"2024-01-22T16:09:47Z","abstract_excerpt":"Detecting text generated by modern large language models is thought to be hard, as both LLMs and humans can exhibit a wide range of complex behaviors. However, we find that a score based on contrasting two closely related language models is highly accurate at separating human-generated and machine-generated text. Based on this mechanism, we propose a novel LLM detector that only requires simple calculations using a pair of pre-trained LLMs. The method, called Binoculars, achieves state-of-the-art accuracy without any training data. It is capable of spotting machine text from a range of modern "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2401.12070","kind":"arxiv","version":3},"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/2401.12070/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":"2401.12070","created_at":"2026-07-05T09:19:53.649683+00:00"},{"alias_kind":"arxiv_version","alias_value":"2401.12070v3","created_at":"2026-07-05T09:19:53.649683+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2401.12070","created_at":"2026-07-05T09:19:53.649683+00:00"},{"alias_kind":"pith_short_12","alias_value":"7SNFBXEIXODA","created_at":"2026-07-05T09:19:53.649683+00:00"},{"alias_kind":"pith_short_16","alias_value":"7SNFBXEIXODAJYKX","created_at":"2026-07-05T09:19:53.649683+00:00"},{"alias_kind":"pith_short_8","alias_value":"7SNFBXEI","created_at":"2026-07-05T09:19:53.649683+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":18,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.25152","citing_title":"Hitting a Moving Target: Test-Time Adaptation for AI Text Detection under Continual Distribution Shift","ref_index":26,"is_internal_anchor":false},{"citing_arxiv_id":"2606.22454","citing_title":"CASPER in the Machine: Insights into Character Variety in LLM-Generated Stories","ref_index":6,"is_internal_anchor":false},{"citing_arxiv_id":"2606.13227","citing_title":"PolyAlign: Conditional Human-Distribution Alignment","ref_index":50,"is_internal_anchor":false},{"citing_arxiv_id":"2606.04205","citing_title":"DetectZoo: A Unified Toolkit for AI-Generated Content Detection Across Text, Audio, and Image Modalities","ref_index":10,"is_internal_anchor":false},{"citing_arxiv_id":"2606.03330","citing_title":"FLIPS: Instance-Fingerprinting for LLMs via Pseudo-random Sequences","ref_index":23,"is_internal_anchor":false},{"citing_arxiv_id":"2605.12890","citing_title":"Steer-to-Detect: Probing Hidden Representations for Detection of LLM-Generated Texts","ref_index":21,"is_internal_anchor":false},{"citing_arxiv_id":"2605.20761","citing_title":"Findings of the Counter Turing Test: AI-Generated Text Detection","ref_index":13,"is_internal_anchor":false},{"citing_arxiv_id":"2606.00392","citing_title":"Detector-Evasive LLM Paraphrasing via Constrained Policy Optimization","ref_index":27,"is_internal_anchor":false},{"citing_arxiv_id":"2606.00545","citing_title":"The Assistant as a Privileged Persona: A canonical reference in cross-persona self-recognition","ref_index":9,"is_internal_anchor":false},{"citing_arxiv_id":"2606.07313","citing_title":"SV-Detect: AI-generated Text Detection with Steering Vectors","ref_index":9,"is_internal_anchor":false},{"citing_arxiv_id":"2502.11336","citing_title":"ExaGPT: Example-Based Machine-Generated Text Detection for Human Interpretability","ref_index":12,"is_internal_anchor":false},{"citing_arxiv_id":"2605.22654","citing_title":"Seeing the Poem: Image-Semantic Detection of AI-Generated Modern Chinese Poetry with MLLMs","ref_index":107,"is_internal_anchor":false},{"citing_arxiv_id":"2605.20761","citing_title":"Findings of the Counter Turing Test: AI-Generated Text Detection","ref_index":10,"is_internal_anchor":false},{"citing_arxiv_id":"2605.16471","citing_title":"From AI-Generated Content to Agentic Action: Security and Safety Threats in Generative AI","ref_index":52,"is_internal_anchor":false},{"citing_arxiv_id":"2605.19516","citing_title":"Base Models Look Human To AI Detectors","ref_index":12,"is_internal_anchor":false},{"citing_arxiv_id":"2509.22055","citing_title":"RedNote-Vibe: A Dataset for Capturing Temporal Dynamics of AI-Generated Text in Lifestyle Social Media","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2605.14240","citing_title":"Paraphrasing Attack Resilience of Various AI-Generated Text Detection Methods","ref_index":5,"is_internal_anchor":false},{"citing_arxiv_id":"2604.26965","citing_title":"The Impact of AI-Generated Text on the Internet","ref_index":13,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/7SNFBXEIXODAJYKXKQM2UTNL3H","json":"https://pith.science/pith/7SNFBXEIXODAJYKXKQM2UTNL3H.json","graph_json":"https://pith.science/api/pith-number/7SNFBXEIXODAJYKXKQM2UTNL3H/graph.json","events_json":"https://pith.science/api/pith-number/7SNFBXEIXODAJYKXKQM2UTNL3H/events.json","paper":"https://pith.science/paper/7SNFBXEI"},"agent_actions":{"view_html":"https://pith.science/pith/7SNFBXEIXODAJYKXKQM2UTNL3H","download_json":"https://pith.science/pith/7SNFBXEIXODAJYKXKQM2UTNL3H.json","view_paper":"https://pith.science/paper/7SNFBXEI","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2401.12070&json=true","fetch_graph":"https://pith.science/api/pith-number/7SNFBXEIXODAJYKXKQM2UTNL3H/graph.json","fetch_events":"https://pith.science/api/pith-number/7SNFBXEIXODAJYKXKQM2UTNL3H/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/7SNFBXEIXODAJYKXKQM2UTNL3H/action/timestamp_anchor","attest_storage":"https://pith.science/pith/7SNFBXEIXODAJYKXKQM2UTNL3H/action/storage_attestation","attest_author":"https://pith.science/pith/7SNFBXEIXODAJYKXKQM2UTNL3H/action/author_attestation","sign_citation":"https://pith.science/pith/7SNFBXEIXODAJYKXKQM2UTNL3H/action/citation_signature","submit_replication":"https://pith.science/pith/7SNFBXEIXODAJYKXKQM2UTNL3H/action/replication_record"}},"created_at":"2026-07-05T09:19:53.649683+00:00","updated_at":"2026-07-05T09:19:53.649683+00:00"}