{"id":"W6920336244","doi":"10.60692/mf63k-29160","title":"Hash data files for \"Hashes are not suitable to verify fixity of the public archived web\"","year":2022,"lang":"en","type":"article","venue":"Greater South Information System","topic":"Web Data Mining and Analysis","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Columbia College","funders":"","keywords":"Hash function; SHA-2; Merkle tree; Hash tree; Construct (python library); Set (abstract data type); Hash chain; Double hashing","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.007943024,0.0007058547,0.0007686056,0.002738648,0.001291784,0.002207627,0.001270331,0.001203859,0.005425587],"category_scores_gemma":[0.08793563,0.0005829259,0.0004595102,0.003830649,0.001774254,0.006525818,0.002202851,0.001551554,0.00359604],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001060807,"about_ca_system_score_gemma":0.001339277,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002267043,"about_ca_topic_score_gemma":0.002453114,"domain_scores_codex":[0.9921358,0.001360199,0.001275959,0.001676642,0.003044594,0.000506982],"domain_scores_gemma":[0.900319,0.04316417,0.01041847,0.03423686,0.01089249,0.0009689877],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"not_applicable","study_design_scores_codex":[0.005712825,0.0007839781,0.4538381,0.004005761,0.0008589893,0.001228391,0.006827432,0.008602229,0.02966438,0.01974475,0.1810075,0.2877258],"study_design_scores_gemma":[0.0008450692,0.001821602,0.3892008,0.001151303,0.0004019649,0.004109356,0.006612633,0.07825161,0.1294538,0.03142828,0.3562232,0.0005003416],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"dataset","genre_scores_codex":[0.7973173,0.001333914,0.04825997,0.001452424,0.001009116,0.0006240903,0.1024678,0.03800178,0.009533605],"genre_scores_gemma":[0.8559065,0.0003222578,0.05313171,0.000546808,0.0001955488,0.0006607555,0.0813833,0.003603431,0.004249682],"genre_candidate":"dataset","genre_consensus":null,"teacher_disagreement_score":0.007943024,"threshold_uncertainty_score":0.04200721,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1013182629852221,"score_gpt":0.2350507976600053,"score_spread":0.1337325346747832,"validation_status":"score_only:v0-immature-baseline","note":"Baseline scores from an immature model (maturity gate not passed). Scores rank; they never assert a category."}}