Hugging Face's logo

rpeel
/
glitext-multilingual-small

ONNX
glitext
Model card Files
xet
 Community
glitext-multilingual-small / modelaudit.json
rpeel's picture
rpeel
Update model card and security scan results
c3aa9e0 verified
raw
history blame contribute delete
113 kB
{
 "tool": "modelaudit",
 "tool_version": "0.2.40",
 "scanned_at": "2026-04-27T00:57:11Z",
 "files": {
 "model.onnx": {
 "size_mb": 688.8,
 "sha256": "6b349feedf138431e22be9bc26221511e2774503517c8713c0531bee499d4164"
 }
 },
 "audit": {
 "bytes_scanned": 709454707,
 "issues": [
 {
 "message": "Weight distribution analysis skipped one or more eligible ONNX initializers",
 "severity": "info",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/model.onnx",
 "details": {
 "scan_outcome_reason": "onnx_weight_distribution_analysis_incomplete",
 "coverage_gap": "partial_initializer_coverage",
 "eligible_initializers": 70,
 "analyzed_initializers": 69,
 "external_initializers_skipped": 0,
 "oversized_initializers_skipped": 1,
 "extraction_failures": 0,
 "max_array_size": 104857600
 },
 "timestamp": 1777251422.4721746,
 "type": "onnx_check",
 "rule_code": "S902"
 },
 {
 "message": "Layer 'core.token_rep_layer.bert_layer.model.encoder.block.0.layer.0.SelfAttention.relative_attention_bias.weight' output neuron 1 has unusually dissimilar weights",
 "severity": "info",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/model.onnx",
 "details": {
 "layer": "core.token_rep_layer.bert_layer.model.encoder.block.0.layer.0.SelfAttention.relative_attention_bias.weight",
 "neuron_index": 1,
 "max_similarity_to_others": 0.6993791460990906,
 "weight_norm": 67.66498565673828,
 "total_outputs": 6,
 "analysis_method": "structural_analysis"
 },
 "why": "Neurons with weight patterns completely unlike others in the same layer are uncommon in standard training. This dissimilarity (measured by cosine similarity below threshold) may indicate injected functionality or training irregularities.",
 "timestamp": 1777251424.5326557,
 "type": "onnx_check",
 "rule_code": "S803"
 },
 {
 "message": "Layer 'core.token_rep_layer.bert_layer.model.encoder.block.0.layer.0.SelfAttention.relative_attention_bias.weight' output neuron 2 has unusually dissimilar weights",
 "severity": "info",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/model.onnx",
 "details": {
 "layer": "core.token_rep_layer.bert_layer.model.encoder.block.0.layer.0.SelfAttention.relative_attention_bias.weight",
 "neuron_index": 2,
 "max_similarity_to_others": 0.6993791460990906,
 "weight_norm": 40.88163757324219,
 "total_outputs": 6,
 "analysis_method": "structural_analysis"
 },
 "why": "Neurons with weight patterns completely unlike others in the same layer are uncommon in standard training. This dissimilarity (measured by cosine similarity below threshold) may indicate injected functionality or training irregularities.",
 "timestamp": 1777251424.5335255,
 "type": "onnx_check",
 "rule_code": "S803"
 },
 {
 "message": "Layer 'core.token_rep_layer.bert_layer.model.encoder.block.0.layer.0.SelfAttention.relative_attention_bias.weight' output neuron 4 has unusually dissimilar weights",
 "severity": "info",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/model.onnx",
 "details": {
 "layer": "core.token_rep_layer.bert_layer.model.encoder.block.0.layer.0.SelfAttention.relative_attention_bias.weight",
 "neuron_index": 4,
 "max_similarity_to_others": 0.37672263383865356,
 "weight_norm": 35.123008728027344,
 "total_outputs": 6,
 "analysis_method": "structural_analysis"
 },
 "why": "Neurons with weight patterns completely unlike others in the same layer are uncommon in standard training. This dissimilarity (measured by cosine similarity below threshold) may indicate injected functionality or training irregularities.",
 "timestamp": 1777251424.534183,
 "type": "onnx_check",
 "rule_code": "S803"
 },
 {
 "message": "Layer 'core.token_rep_layer.bert_layer.model.encoder.block.0.layer.0.SelfAttention.relative_attention_bias.weight' has neurons with extremely large weight values",
 "severity": "info",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/model.onnx",
 "details": {
 "layer": "core.token_rep_layer.bert_layer.model.encoder.block.0.layer.0.SelfAttention.relative_attention_bias.weight",
 "affected_neurons": [
 1
 ],
 "total_affected": 1,
 "num_extreme_weights": 5,
 "threshold": 17.915985107421875,
 "max_weight": 21.875,
 "total_outputs": 6,
 "analysis_method": "structural_analysis"
 },
 "why": "Weight values that are orders of magnitude larger than typical can cause numerical instability, overflow attacks, or may encode hidden data. Detection uses statistical analysis rather than name-based classification to avoid security bypasses.",
 "timestamp": 1777251424.5348,
 "type": "onnx_check",
 "rule_code": "S802"
 },
 {
 "message": "Layer 'onnx::MatMul_1495' has 1 output neurons with abnormal weight magnitudes",
 "severity": "info",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/model.onnx",
 "details": {
 "layer": "onnx::MatMul_1495",
 "outlier_neurons": [
 285
 ],
 "total_outliers": 1,
 "outlier_percentage": 0.26041666666666663,
 "z_scores": [
 3.5612642765045166
 ],
 "weight_norms": [
 5.0775933265686035
 ],
 "mean_norm": 9.226380348205566,
 "std_norm": 1.164975881576538,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 "why": "Neurons with weight magnitudes significantly different from others in the same layer may indicate tampering, backdoors, or training anomalies. These outliers are flagged when their statistical z-score exceeds the threshold. Thresholds are adjusted based on structural analysis of the model architecture.",
 "timestamp": 1777251424.5353432,
 "type": "onnx_check"
 },
 {
 "message": "Layer 'onnx::MatMul_1499' has 4 output neurons with abnormal weight magnitudes",
 "severity": "info",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/model.onnx",
 "details": {
 "layer": "onnx::MatMul_1499",
 "outlier_neurons": [
 26,
 106,
 334,
 420
 ],
 "total_outliers": 4,
 "outlier_percentage": 0.78125,
 "z_scores": [
 15.07185173034668,
 9.374092102050781,
 11.908661842346191,
 4.552164554595947
 ],
 "weight_norms": [
 47.338375091552734,
 32.29711151123047,
 38.98801040649414,
 19.567920684814453
 ],
 "mean_norm": 7.550865173339844,
 "std_norm": 2.63985538482666,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 "why": "Neurons with weight magnitudes significantly different from others in the same layer may indicate tampering, backdoors, or training anomalies. These outliers are flagged when their statistical z-score exceeds the threshold. Thresholds are adjusted based on structural analysis of the model architecture.",
 "timestamp": 1777251424.5358653,
 "type": "onnx_check"
 },
 {
 "message": "Layer 'onnx::MatMul_1501' has 9 output neurons with abnormal weight magnitudes",
 "severity": "info",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/model.onnx",
 "details": {
 "layer": "onnx::MatMul_1501",
 "outlier_neurons": [
 148,
 176,
 255,
 291,
 641,
 738,
 819,
 991,
 1016
 ],
 "total_outliers": 9,
 "outlier_percentage": 0.87890625,
 "z_scores": [
 3.5479860305786133,
 3.317765951156616,
 8.566732406616211,
 3.050501823425293,
 3.4581427574157715,
 3.5891425609588623,
 3.3659801483154297,
 3.3039772510528564,
 3.0366768836975098
 ],
 "weight_norms": [
 4.05752420425415,
 12.566008567810059,
 19.07086753845215,
 4.674038887023926,
 4.168863773345947,
 12.902316093444824,
 4.283077716827393,
 12.548920631408691,
 4.691171646118164
 ],
 "mean_norm": 8.454418182373047,
 "std_norm": 1.239264726638794,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 "why": "Neurons with weight magnitudes significantly different from others in the same layer may indicate tampering, backdoors, or training anomalies. These outliers are flagged when their statistical z-score exceeds the threshold. Thresholds are adjusted based on structural analysis of the model architecture.",
 "timestamp": 1777251424.5363708,
 "type": "onnx_check"
 },
 {
 "message": "Layer 'onnx::MatMul_1502' has 4 output neurons with abnormal weight magnitudes",
 "severity": "info",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/model.onnx",
 "details": {
 "layer": "onnx::MatMul_1502",
 "outlier_neurons": [
 26,
 106,
 334,
 420
 ],
 "total_outliers": 4,
 "outlier_percentage": 0.78125,
 "z_scores": [
 16.296117782592773,
 8.717620849609375,
 7.3806352615356445,
 8.0861234664917
 ],
 "weight_norms": [
 31.29328155517578,
 20.282611846923828,
 18.34012794494629,
 19.36511993408203
 ],
 "mean_norm": 7.616927623748779,
 "std_norm": 1.452883005142212,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 "why": "Neurons with weight magnitudes significantly different from others in the same layer may indicate tampering, backdoors, or training anomalies. These outliers are flagged when their statistical z-score exceeds the threshold. Thresholds are adjusted based on structural analysis of the model architecture.",
 "timestamp": 1777251424.5369072,
 "type": "onnx_check"
 },
 {
 "message": "Layer 'onnx::MatMul_1517' has 4 output neurons with abnormal weight magnitudes",
 "severity": "info",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/model.onnx",
 "details": {
 "layer": "onnx::MatMul_1517",
 "outlier_neurons": [
 26,
 106,
 334,
 420
 ],
 "total_outliers": 4,
 "outlier_percentage": 0.78125,
 "z_scores": [
 7.922232151031494,
 13.225406646728516,
 12.309375762939453,
 7.317721366882324
 ],
 "weight_norms": [
 26.131452560424805,
 37.51491165161133,
 35.54861831665039,
 24.83384895324707
 ],
 "mean_norm": 9.126093864440918,
 "std_norm": 2.146536350250244,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 "why": "Neurons with weight magnitudes significantly different from others in the same layer may indicate tampering, backdoors, or training anomalies. These outliers are flagged when their statistical z-score exceeds the threshold. Thresholds are adjusted based on structural analysis of the model architecture.",
 "timestamp": 1777251424.537414,
 "type": "onnx_check"
 },
 {
 "message": "Layer 'onnx::MatMul_1518' has 8 output neurons with abnormal weight magnitudes",
 "severity": "info",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/model.onnx",
 "details": {
 "layer": "onnx::MatMul_1518",
 "outlier_neurons": [
 30,
 77,
 184,
 425,
 668,
 735,
 761,
 985
 ],
 "total_outliers": 8,
 "outlier_percentage": 0.78125,
 "z_scores": [
 3.8372180461883545,
 5.266956806182861,
 3.266873836517334,
 3.6158640384674072,
 3.9202046394348145,
 4.8188629150390625,
 4.044105052947998,
 3.048353910446167
 ],
 "weight_norms": [
 11.20538330078125,
 12.909429550170898,
 10.525613784790039,
 10.941560745239258,
 11.304291725158691,
 12.375365257263184,
 11.451963424682617,
 10.265169143676758
 ],
 "mean_norm": 6.63196325302124,
 "std_norm": 1.1918582916259766,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 "why": "Neurons with weight magnitudes significantly different from others in the same layer may indicate tampering, backdoors, or training anomalies. These outliers are flagged when their statistical z-score exceeds the threshold. Thresholds are adjusted based on structural analysis of the model architecture.",
 "timestamp": 1777251424.5379314,
 "type": "onnx_check"
 },
 {
 "message": "Layer 'onnx::MatMul_1519' has 10 output neurons with abnormal weight magnitudes",
 "severity": "info",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/model.onnx",
 "details": {
 "layer": "onnx::MatMul_1519",
 "outlier_neurons": [
 139,
 191,
 231,
 287,
 300,
 301,
 664,
 749,
 801,
 1023
 ],
 "total_outliers": 10,
 "outlier_percentage": 0.9765625,
 "z_scores": [
 3.0819313526153564,
 3.734606981277466,
 3.059579372406006,
 3.0208399295806885,
 3.091632843017578,
 4.155886650085449,
 3.211397171020508,
 3.5522100925445557,
 3.0440330505371094,
 3.232992649078369
 ],
 "weight_norms": [
 14.858200073242188,
 15.806731224060059,
 14.825716018676758,
 14.769415855407715,
 5.886178970336914,
 16.418975830078125,
 15.04635238647461,
 15.541654586791992,
 14.803122520446777,
 15.077736854553223
 ],
 "mean_norm": 10.379239082336426,
 "std_norm": 1.4532967805862427,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 "why": "Neurons with weight magnitudes significantly different from others in the same layer may indicate tampering, backdoors, or training anomalies. These outliers are flagged when their statistical z-score exceeds the threshold. Thresholds are adjusted based on structural analysis of the model architecture.",
 "timestamp": 1777251424.5384352,
 "type": "onnx_check"
 },
 {
 "message": "Layer 'onnx::MatMul_1520' has 4 output neurons with abnormal weight magnitudes",
 "severity": "info",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/model.onnx",
 "details": {
 "layer": "onnx::MatMul_1520",
 "outlier_neurons": [
 26,
 106,
 334,
 420
 ],
 "total_outliers": 4,
 "outlier_percentage": 0.78125,
 "z_scores": [
 10.825921058654785,
 10.923227310180664,
 9.819406509399414,
 10.719425201416016
 ],
 "weight_norms": [
 26.44086265563965,
 26.592884063720703,
 24.86839485168457,
 26.274484634399414
 ],
 "mean_norm": 9.52762508392334,
 "std_norm": 1.562290906906128,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 "why": "Neurons with weight magnitudes significantly different from others in the same layer may indicate tampering, backdoors, or training anomalies. These outliers are flagged when their statistical z-score exceeds the threshold. Thresholds are adjusted based on structural analysis of the model architecture.",
 "timestamp": 1777251424.5389445,
 "type": "onnx_check"
 },
 {
 "message": "Layer 'onnx::MatMul_1532' has 1 output neurons with abnormal weight magnitudes",
 "severity": "info",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/model.onnx",
 "details": {
 "layer": "onnx::MatMul_1532",
 "outlier_neurons": [
 24
 ],
 "total_outliers": 1,
 "outlier_percentage": 0.26041666666666663,
 "z_scores": [
 3.5400595664978027
 ],
 "weight_norms": [
 13.14366340637207
 ],
 "mean_norm": 8.88523006439209,
 "std_norm": 1.2029269933700562,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 "why": "Neurons with weight magnitudes significantly different from others in the same layer may indicate tampering, backdoors, or training anomalies. These outliers are flagged when their statistical z-score exceeds the threshold. Thresholds are adjusted based on structural analysis of the model architecture.",
 "timestamp": 1777251424.5394351,
 "type": "onnx_check"
 },
 {
 "message": "Layer 'onnx::MatMul_1535' has 4 output neurons with abnormal weight magnitudes",
 "severity": "info",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/model.onnx",
 "details": {
 "layer": "onnx::MatMul_1535",
 "outlier_neurons": [
 26,
 106,
 334,
 420
 ],
 "total_outliers": 4,
 "outlier_percentage": 0.78125,
 "z_scores": [
 4.199339866638184,
 6.5627593994140625,
 17.913204193115234,
 7.438827037811279
 ],
 "weight_norms": [
 19.08266258239746,
 24.46711540222168,
 50.326229095458984,
 26.463014602661133
 ],
 "mean_norm": 9.51552963256836,
 "std_norm": 2.2782468795776367,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 "why": "Neurons with weight magnitudes significantly different from others in the same layer may indicate tampering, backdoors, or training anomalies. These outliers are flagged when their statistical z-score exceeds the threshold. Thresholds are adjusted based on structural analysis of the model architecture.",
 "timestamp": 1777251424.5399582,
 "type": "onnx_check"
 },
 {
 "message": "Layer 'onnx::MatMul_1537' has 4 output neurons with abnormal weight magnitudes",
 "severity": "info",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/model.onnx",
 "details": {
 "layer": "onnx::MatMul_1537",
 "outlier_neurons": [
 443,
 468,
 721,
 867
 ],
 "total_outliers": 4,
 "outlier_percentage": 0.390625,
 "z_scores": [
 3.197878360748291,
 3.6349661350250244,
 3.3462107181549072,
 17.198549270629883
 ],
 "weight_norms": [
 21.443992614746094,
 22.682941436767578,
 21.86444854736328,
 61.1296501159668
 ],
 "mean_norm": 12.379434585571289,
 "std_norm": 2.8345537185668945,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 "why": "Neurons with weight magnitudes significantly different from others in the same layer may indicate tampering, backdoors, or training anomalies. These outliers are flagged when their statistical z-score exceeds the threshold. Thresholds are adjusted based on structural analysis of the model architecture.",
 "timestamp": 1777251424.5404491,
 "type": "onnx_check"
 },
 {
 "message": "Layer 'onnx::MatMul_1538' has 4 output neurons with abnormal weight magnitudes",
 "severity": "info",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/model.onnx",
 "details": {
 "layer": "onnx::MatMul_1538",
 "outlier_neurons": [
 26,
 106,
 334,
 420
 ],
 "total_outliers": 4,
 "outlier_percentage": 0.78125,
 "z_scores": [
 7.756042003631592,
 7.21874475479126,
 15.043734550476074,
 10.037714958190918
 ],
 "weight_norms": [
 22.36002540588379,
 21.563430786132812,
 33.164730072021484,
 25.74282455444336
 ],
 "mean_norm": 10.860949516296387,
 "std_norm": 1.4825959205627441,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 "why": "Neurons with weight magnitudes significantly different from others in the same layer may indicate tampering, backdoors, or training anomalies. These outliers are flagged when their statistical z-score exceeds the threshold. Thresholds are adjusted based on structural analysis of the model architecture.",
 "timestamp": 1777251424.5409513,
 "type": "onnx_check"
 },
 {
 "message": "Layer 'onnx::MatMul_1553' has 4 output neurons with abnormal weight magnitudes",
 "severity": "info",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/model.onnx",
 "details": {
 "layer": "onnx::MatMul_1553",
 "outlier_neurons": [
 26,
 106,
 334,
 420
 ],
 "total_outliers": 4,
 "outlier_percentage": 0.78125,
 "z_scores": [
 3.432584285736084,
 5.6761674880981445,
 17.938501358032227,
 7.101133346557617
 ],
 "weight_norms": [
 16.69535255432129,
 20.915910720825195,
 43.983428955078125,
 23.596511840820312
 ],
 "mean_norm": 10.238082885742188,
 "std_norm": 1.8811686038970947,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 "why": "Neurons with weight magnitudes significantly different from others in the same layer may indicate tampering, backdoors, or training anomalies. These outliers are flagged when their statistical z-score exceeds the threshold. Thresholds are adjusted based on structural analysis of the model architecture.",
 "timestamp": 1777251424.5414436,
 "type": "onnx_check"
 },
 {
 "message": "Layer 'onnx::MatMul_1556' has 4 output neurons with abnormal weight magnitudes",
 "severity": "info",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/model.onnx",
 "details": {
 "layer": "onnx::MatMul_1556",
 "outlier_neurons": [
 26,
 106,
 334,
 420
 ],
 "total_outliers": 4,
 "outlier_percentage": 0.78125,
 "z_scores": [
 8.058356285095215,
 7.137109756469727,
 15.326730728149414,
 10.435029029846191
 ],
 "weight_norms": [
 25.194787979125977,
 23.626604080200195,
 37.567317962646484,
 29.240459442138672
 ],
 "mean_norm": 11.477519035339355,
 "std_norm": 1.7022415399551392,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 "why": "Neurons with weight magnitudes significantly different from others in the same layer may indicate tampering, backdoors, or training anomalies. These outliers are flagged when their statistical z-score exceeds the threshold. Thresholds are adjusted based on structural analysis of the model architecture.",
 "timestamp": 1777251424.5419538,
 "type": "onnx_check"
 },
 {
 "message": "Layer 'onnx::MatMul_1557' has 1 output neurons with abnormal weight magnitudes",
 "severity": "info",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/model.onnx",
 "details": {
 "layer": "onnx::MatMul_1557",
 "outlier_neurons": [
 223
 ],
 "total_outliers": 1,
 "outlier_percentage": 0.26041666666666663,
 "z_scores": [
 3.4946324825286865
 ],
 "weight_norms": [
 0.9592359066009521
 ],
 "mean_norm": 0.6574398875236511,
 "std_norm": 0.08635987341403961,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 "why": "Neurons with weight magnitudes significantly different from others in the same layer may indicate tampering, backdoors, or training anomalies. These outliers are flagged when their statistical z-score exceeds the threshold. Thresholds are adjusted based on structural analysis of the model architecture.",
 "timestamp": 1777251424.542449,
 "type": "onnx_check"
 },
 {
 "message": "Layer 'onnx::MatMul_1567' has 2 output neurons with abnormal weight magnitudes",
 "severity": "info",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/model.onnx",
 "details": {
 "layer": "onnx::MatMul_1567",
 "outlier_neurons": [
 223,
 252
 ],
 "total_outliers": 2,
 "outlier_percentage": 0.5208333333333333,
 "z_scores": [
 3.2663497924804688,
 3.0496039390563965
 ],
 "weight_norms": [
 8.01538372039795,
 7.8404107093811035
 ],
 "mean_norm": 5.378547668457031,
 "std_norm": 0.8072730302810669,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 "why": "Neurons with weight magnitudes significantly different from others in the same layer may indicate tampering, backdoors, or training anomalies. These outliers are flagged when their statistical z-score exceeds the threshold. Thresholds are adjusted based on structural analysis of the model architecture.",
 "timestamp": 1777251424.5429506,
 "type": "onnx_check"
 },
 {
 "message": "Layer 'onnx::MatMul_1568' has 1 output neurons with abnormal weight magnitudes",
 "severity": "info",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/model.onnx",
 "details": {
 "layer": "onnx::MatMul_1568",
 "outlier_neurons": [
 12
 ],
 "total_outliers": 1,
 "outlier_percentage": 0.26041666666666663,
 "z_scores": [
 3.0931005477905273
 ],
 "weight_norms": [
 9.757271766662598
 ],
 "mean_norm": 13.355530738830566,
 "std_norm": 1.1633177995681763,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 "why": "Neurons with weight magnitudes significantly different from others in the same layer may indicate tampering, backdoors, or training anomalies. These outliers are flagged when their statistical z-score exceeds the threshold. Thresholds are adjusted based on structural analysis of the model architecture.",
 "timestamp": 1777251424.543441,
 "type": "onnx_check"
 },
 {
 "message": "Layer 'onnx::MatMul_1571' has 3 output neurons with abnormal weight magnitudes",
 "severity": "info",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/model.onnx",
 "details": {
 "layer": "onnx::MatMul_1571",
 "outlier_neurons": [
 106,
 334,
 420
 ],
 "total_outliers": 3,
 "outlier_percentage": 0.5859375,
 "z_scores": [
 3.732685089111328,
 18.809659957885742,
 3.911799669265747
 ],
 "weight_norms": [
 23.456144332885742,
 66.8714370727539,
 23.9719181060791
 ],
 "mean_norm": 12.707594871520996,
 "std_norm": 2.879575729370117,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 "why": "Neurons with weight magnitudes significantly different from others in the same layer may indicate tampering, backdoors, or training anomalies. These outliers are flagged when their statistical z-score exceeds the threshold. Thresholds are adjusted based on structural analysis of the model architecture.",
 "timestamp": 1777251424.5439544,
 "type": "onnx_check"
 },
 {
 "message": "Layer 'onnx::MatMul_1573' has 3 output neurons with abnormal weight magnitudes",
 "severity": "info",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/model.onnx",
 "details": {
 "layer": "onnx::MatMul_1573",
 "outlier_neurons": [
 381,
 468,
 519
 ],
 "total_outliers": 3,
 "outlier_percentage": 0.29296875,
 "z_scores": [
 4.015261173248291,
 3.3636810779571533,
 3.3158137798309326
 ],
 "weight_norms": [
 26.825809478759766,
 24.75054359436035,
 24.598087310791016
 ],
 "mean_norm": 14.037303924560547,
 "std_norm": 3.1849749088287354,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 "why": "Neurons with weight magnitudes significantly different from others in the same layer may indicate tampering, backdoors, or training anomalies. These outliers are flagged when their statistical z-score exceeds the threshold. Thresholds are adjusted based on structural analysis of the model architecture.",
 "timestamp": 1777251424.5444496,
 "type": "onnx_check"
 },
 {
 "message": "Layer 'onnx::MatMul_1574' has 4 output neurons with abnormal weight magnitudes",
 "severity": "info",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/model.onnx",
 "details": {
 "layer": "onnx::MatMul_1574",
 "outlier_neurons": [
 26,
 106,
 334,
 420
 ],
 "total_outliers": 4,
 "outlier_percentage": 0.78125,
 "z_scores": [
 5.549093723297119,
 6.461915016174316,
 17.216140747070312,
 8.640796661376953
 ],
 "weight_norms": [
 23.100589752197266,
 24.974454879760742,
 47.051029205322266,
 29.447324752807617
 ],
 "mean_norm": 11.709256172180176,
 "std_norm": 2.052827835083008,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 "why": "Neurons with weight magnitudes significantly different from others in the same layer may indicate tampering, backdoors, or training anomalies. These outliers are flagged when their statistical z-score exceeds the threshold. Thresholds are adjusted based on structural analysis of the model architecture.",
 "timestamp": 1777251424.544953,
 "type": "onnx_check"
 },
 {
 "message": "Layer 'onnx::MatMul_1589' has 2 output neurons with abnormal weight magnitudes",
 "severity": "info",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/model.onnx",
 "details": {
 "layer": "onnx::MatMul_1589",
 "outlier_neurons": [
 106,
 334
 ],
 "total_outliers": 2,
 "outlier_percentage": 0.390625,
 "z_scores": [
 3.0379512310028076,
 17.739160537719727
 ],
 "weight_norms": [
 22.80105972290039,
 68.33921813964844
 ],
 "mean_norm": 13.390766143798828,
 "std_norm": 3.097579002380371,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 "why": "Neurons with weight magnitudes significantly different from others in the same layer may indicate tampering, backdoors, or training anomalies. These outliers are flagged when their statistical z-score exceeds the threshold. Thresholds are adjusted based on structural analysis of the model architecture.",
 "timestamp": 1777251424.545445,
 "type": "onnx_check"
 },
 {
 "message": "Layer 'onnx::MatMul_1591' has 9 output neurons with abnormal weight magnitudes",
 "severity": "info",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/model.onnx",
 "details": {
 "layer": "onnx::MatMul_1591",
 "outlier_neurons": [
 57,
 119,
 288,
 291,
 457,
 812,
 827,
 915,
 922
 ],
 "total_outliers": 9,
 "outlier_percentage": 0.87890625,
 "z_scores": [
 3.6402089595794678,
 3.06794810295105,
 3.236544370651245,
 3.8063037395477295,
 3.7410624027252197,
 4.355141639709473,
 3.2011234760284424,
 3.1648313999176025,
 3.655369281768799
 ],
 "weight_norms": [
 27.504365921020508,
 25.553974151611328,
 26.12858772277832,
 28.070453643798828,
 27.84809684753418,
 29.941015243530273,
 26.00786590576172,
 25.884174346923828,
 27.5560359954834
 ],
 "mean_norm": 15.09772777557373,
 "std_norm": 3.408221483230591,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 "why": "Neurons with weight magnitudes significantly different from others in the same layer may indicate tampering, backdoors, or training anomalies. These outliers are flagged when their statistical z-score exceeds the threshold. Thresholds are adjusted based on structural analysis of the model architecture.",
 "timestamp": 1777251424.545947,
 "type": "onnx_check"
 },
 {
 "message": "Layer 'onnx::MatMul_1592' has 5 output neurons with abnormal weight magnitudes",
 "severity": "info",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/model.onnx",
 "details": {
 "layer": "onnx::MatMul_1592",
 "outlier_neurons": [
 26,
 106,
 334,
 420,
 448
 ],
 "total_outliers": 5,
 "outlier_percentage": 0.9765625,
 "z_scores": [
 6.068371295928955,
 5.392306327819824,
 16.087129592895508,
 8.780937194824219,
 3.143547296524048
 ],
 "weight_norms": [
 25.586109161376953,
 24.145708084106445,
 46.93174362182617,
 31.365413665771484,
 19.354576110839844
 ],
 "mean_norm": 12.657038688659668,
 "std_norm": 2.1305668354034424,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 "why": "Neurons with weight magnitudes significantly different from others in the same layer may indicate tampering, backdoors, or training anomalies. These outliers are flagged when their statistical z-score exceeds the threshold. Thresholds are adjusted based on structural analysis of the model architecture.",
 "timestamp": 1777251424.546437,
 "type": "onnx_check"
 },
 {
 "message": "Layer 'onnx::MatMul_1604' has 1 output neurons with abnormal weight magnitudes",
 "severity": "info",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/model.onnx",
 "details": {
 "layer": "onnx::MatMul_1604",
 "outlier_neurons": [
 250
 ],
 "total_outliers": 1,
 "outlier_percentage": 0.26041666666666663,
 "z_scores": [
 3.181227922439575
 ],
 "weight_norms": [
 21.17793083190918
 ],
 "mean_norm": 15.04564380645752,
 "std_norm": 1.9276478290557861,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 "why": "Neurons with weight magnitudes significantly different from others in the same layer may indicate tampering, backdoors, or training anomalies. These outliers are flagged when their statistical z-score exceeds the threshold. Thresholds are adjusted based on structural analysis of the model architecture.",
 "timestamp": 1777251424.5469692,
 "type": "onnx_check"
 },
 {
 "message": "Layer 'onnx::MatMul_1607' has 2 output neurons with abnormal weight magnitudes",
 "severity": "info",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/model.onnx",
 "details": {
 "layer": "onnx::MatMul_1607",
 "outlier_neurons": [
 26,
 334
 ],
 "total_outliers": 2,
 "outlier_percentage": 0.390625,
 "z_scores": [
 3.0200729370117188,
 15.381612777709961
 ],
 "weight_norms": [
 21.44904899597168,
 54.05788040161133
 ],
 "mean_norm": 13.482318878173828,
 "std_norm": 2.6379263401031494,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 "why": "Neurons with weight magnitudes significantly different from others in the same layer may indicate tampering, backdoors, or training anomalies. These outliers are flagged when their statistical z-score exceeds the threshold. Thresholds are adjusted based on structural analysis of the model architecture.",
 "timestamp": 1777251424.5474613,
 "type": "onnx_check"
 },
 {
 "message": "Layer 'onnx::MatMul_1609' has 9 output neurons with abnormal weight magnitudes",
 "severity": "info",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/model.onnx",
 "details": {
 "layer": "onnx::MatMul_1609",
 "outlier_neurons": [
 23,
 159,
 244,
 310,
 399,
 425,
 819,
 849,
 1002
 ],
 "total_outliers": 9,
 "outlier_percentage": 0.87890625,
 "z_scores": [
 3.364377737045288,
 3.4016523361206055,
 3.229233741760254,
 3.1587109565734863,
 3.251481294631958,
 3.1432199478149414,
 3.158773422241211,
 4.49865198135376,
 3.074720621109009
 ],
 "weight_norms": [
 27.67537498474121,
 27.80193328857422,
 27.216522216796875,
 26.97707748413086,
 27.29205894470215,
 26.924480438232422,
 26.9772891998291,
 31.526561737060547,
 26.691905975341797
 ],
 "mean_norm": 16.25234603881836,
 "std_norm": 3.395287275314331,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 "why": "Neurons with weight magnitudes significantly different from others in the same layer may indicate tampering, backdoors, or training anomalies. These outliers are flagged when their statistical z-score exceeds the threshold. Thresholds are adjusted based on structural analysis of the model architecture.",
 "timestamp": 1777251424.548029,
 "type": "onnx_check"
 },
 {
 "message": "Layer 'onnx::MatMul_1610' has 5 output neurons with abnormal weight magnitudes",
 "severity": "info",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/model.onnx",
 "details": {
 "layer": "onnx::MatMul_1610",
 "outlier_neurons": [
 26,
 106,
 334,
 420,
 448
 ],
 "total_outliers": 5,
 "outlier_percentage": 0.9765625,
 "z_scores": [
 5.163490295410156,
 4.521091938018799,
 12.709145545959473,
 8.433018684387207,
 3.476202964782715
 ],
 "weight_norms": [
 24.45191192626953,
 23.157756805419922,
 39.65314483642578,
 31.038597106933594,
 21.052757263183594
 ],
 "mean_norm": 14.049711227416992,
 "std_norm": 2.0145676136016846,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 "why": "Neurons with weight magnitudes significantly different from others in the same layer may indicate tampering, backdoors, or training anomalies. These outliers are flagged when their statistical z-score exceeds the threshold. Thresholds are adjusted based on structural analysis of the model architecture.",
 "timestamp": 1777251424.5485265,
 "type": "onnx_check"
 },
 {
 "message": "Layer 'onnx::MatMul_1625' has 4 output neurons with abnormal weight magnitudes",
 "severity": "info",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/model.onnx",
 "details": {
 "layer": "onnx::MatMul_1625",
 "outlier_neurons": [
 26,
 106,
 334,
 448
 ],
 "total_outliers": 4,
 "outlier_percentage": 0.78125,
 "z_scores": [
 3.1422200202941895,
 3.2814738750457764,
 14.4247407913208,
 4.578961372375488
 ],
 "weight_norms": [
 28.572736740112305,
 29.122215270996094,
 73.0921859741211,
 34.241943359375
 ],
 "mean_norm": 16.173919677734375,
 "std_norm": 3.945878028869629,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 "why": "Neurons with weight magnitudes significantly different from others in the same layer may indicate tampering, backdoors, or training anomalies. These outliers are flagged when their statistical z-score exceeds the threshold. Thresholds are adjusted based on structural analysis of the model architecture.",
 "timestamp": 1777251424.5490372,
 "type": "onnx_check"
 },
 {
 "message": "Layer 'onnx::MatMul_1628' has 5 output neurons with abnormal weight magnitudes",
 "severity": "info",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/model.onnx",
 "details": {
 "layer": "onnx::MatMul_1628",
 "outlier_neurons": [
 26,
 106,
 334,
 420,
 448
 ],
 "total_outliers": 5,
 "outlier_percentage": 0.9765625,
 "z_scores": [
 6.393590927124023,
 4.238614082336426,
 11.386103630065918,
 7.328854084014893,
 4.357119083404541
 ],
 "weight_norms": [
 35.1436767578125,
 28.690250396728516,
 50.09456253051758,
 37.94447326660156,
 29.04513168334961
 ],
 "mean_norm": 15.997035026550293,
 "std_norm": 2.994661569595337,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 "why": "Neurons with weight magnitudes significantly different from others in the same layer may indicate tampering, backdoors, or training anomalies. These outliers are flagged when their statistical z-score exceeds the threshold. Thresholds are adjusted based on structural analysis of the model architecture.",
 "timestamp": 1777251424.5495365,
 "type": "onnx_check"
 },
 {
 "message": "Layer 'onnx::MatMul_1629' has 1 output neurons with abnormal weight magnitudes",
 "severity": "info",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/model.onnx",
 "details": {
 "layer": "onnx::MatMul_1629",
 "outlier_neurons": [
 61
 ],
 "total_outliers": 1,
 "outlier_percentage": 0.13020833333333331,
 "z_scores": [
 3.446139097213745
 ],
 "weight_norms": [
 0.4687581956386566
 ],
 "mean_norm": 0.5061156153678894,
 "std_norm": 0.010840369388461113,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 "why": "Neurons with weight magnitudes significantly different from others in the same layer may indicate tampering, backdoors, or training anomalies. These outliers are flagged when their statistical z-score exceeds the threshold. Thresholds are adjusted based on structural analysis of the model architecture.",
 "timestamp": 1777251424.5500429,
 "type": "onnx_check"
 },
 {
 "message": "Layer 'onnx::LSTM_1676' has 16 output neurons with abnormal weight magnitudes",
 "severity": "info",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/model.onnx",
 "details": {
 "layer": "onnx::LSTM_1676",
 "outlier_neurons": [
 201,
 246,
 792,
 1035,
 1113,
 1231,
 1440,
 1477,
 1803,
 2121
 ],
 "total_outliers": 16,
 "outlier_percentage": 0.5208333333333333,
 "z_scores": [
 3.188148260116577,
 3.1360602378845215,
 3.598491907119751,
 3.253804922103882,
 3.081761598587036,
 3.066697835922241,
 3.7533764839172363,
 4.241366386413574,
 3.411332130432129,
 3.427668333053589
 ],
 "weight_norms": [
 0.10633181780576706,
 0.1052742451429367,
 0.11466331034898758,
 0.10766489803791046,
 0.10417177528142929,
 0.10386592894792557,
 0.11780804395675659,
 0.1277160495519638,
 0.11086327582597733,
 0.11119496822357178
 ],
 "mean_norm": 0.04160062596201897,
 "std_norm": 0.020303696393966675,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 "why": "Neurons with weight magnitudes significantly different from others in the same layer may indicate tampering, backdoors, or training anomalies. These outliers are flagged when their statistical z-score exceeds the threshold. Thresholds are adjusted based on structural analysis of the model architecture.",
 "timestamp": 1777251424.5505302,
 "type": "onnx_check"
 },
 {
 "message": "Layer 'onnx::MatMul_1679' has 11 output neurons with abnormal weight magnitudes",
 "severity": "info",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/model.onnx",
 "details": {
 "layer": "onnx::MatMul_1679",
 "outlier_neurons": [
 11,
 288,
 1245,
 1488,
 1519,
 1586,
 2048,
 2218,
 2378,
 2629
 ],
 "total_outliers": 11,
 "outlier_percentage": 0.35807291666666663,
 "z_scores": [
 3.2400310039520264,
 3.057278633117676,
 3.505038261413574,
 3.123494863510132,
 3.248295783996582,
 3.3864033222198486,
 3.5900251865386963,
 3.0242512226104736,
 3.2676315307617188,
 3.8947160243988037
 ],
 "weight_norms": [
 0.5666050910949707,
 0.5648508071899414,
 0.501857578754425,
 0.5654864311218262,
 0.5666844248771667,
 0.568010151386261,
 0.5699647665023804,
 0.5064727663993835,
 0.5668700337409973,
 0.5728895664215088
 ],
 "mean_norm": 0.5355032682418823,
 "std_norm": 0.009599235840141773,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 "why": "Neurons with weight magnitudes significantly different from others in the same layer may indicate tampering, backdoors, or training anomalies. These outliers are flagged when their statistical z-score exceeds the threshold. Thresholds are adjusted based on structural analysis of the model architecture.",
 "timestamp": 1777251424.551047,
 "type": "onnx_check"
 },
 {
 "message": "Layer 'onnx::MatMul_1681' has 8 output neurons with abnormal weight magnitudes",
 "severity": "info",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/model.onnx",
 "details": {
 "layer": "onnx::MatMul_1681",
 "outlier_neurons": [
 74,
 415,
 677,
 966,
 2048,
 2753,
 2811,
 2885
 ],
 "total_outliers": 8,
 "outlier_percentage": 0.26041666666666663,
 "z_scores": [
 3.467031478881836,
 3.0368196964263916,
 3.1272900104522705,
 3.0710036754608154,
 3.192030429840088,
 3.0490598678588867,
 3.2949187755584717,
 3.122159242630005
 ],
 "weight_norms": [
 0.5669937133789062,
 0.5629504919052124,
 0.5638007521629333,
 0.5632717609405518,
 0.5644091963768005,
 0.5057541131973267,
 0.5653761625289917,
 0.5637525320053101
 ],
 "mean_norm": 0.5344098210334778,
 "std_norm": 0.009398210793733597,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 "why": "Neurons with weight magnitudes significantly different from others in the same layer may indicate tampering, backdoors, or training anomalies. These outliers are flagged when their statistical z-score exceeds the threshold. Thresholds are adjusted based on structural analysis of the model architecture.",
 "timestamp": 1777251424.551537,
 "type": "onnx_check"
 },
 {
 "message": "Layer 'onnx::MatMul_1690' has 11 output neurons with abnormal weight magnitudes",
 "severity": "info",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/model.onnx",
 "details": {
 "layer": "onnx::MatMul_1690",
 "outlier_neurons": [
 187,
 240,
 922,
 1402,
 1451,
 1695,
 1814,
 2105,
 2374,
 2757
 ],
 "total_outliers": 11,
 "outlier_percentage": 0.35807291666666663,
 "z_scores": [
 3.3429384231567383,
 3.2040178775787354,
 3.137970447540283,
 3.394183874130249,
 3.0068321228027344,
 3.1438193321228027,
 3.0620224475860596,
 3.731614589691162,
 4.197888374328613,
 3.3148562908172607
 ],
 "weight_norms": [
 0.49853774905204773,
 0.5658571124076843,
 0.5651779770851135,
 0.5678125023841858,
 0.5638295412063599,
 0.5652381181716919,
 0.5643970370292664,
 0.571282148361206,
 0.5760766267776489,
 0.5669968128204346
 ],
 "mean_norm": 0.5329116582870483,
 "std_norm": 0.010282543487846851,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 "why": "Neurons with weight magnitudes significantly different from others in the same layer may indicate tampering, backdoors, or training anomalies. These outliers are flagged when their statistical z-score exceeds the threshold. Thresholds are adjusted based on structural analysis of the model architecture.",
 "timestamp": 1777251424.5520506,
 "type": "onnx_check"
 },
 {
 "message": "File does not contain expected XGBoost binary model markers",
 "severity": "info",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/spiece.model",
 "details": {
 "expected_patterns": [
 "binf",
 "gbtree",
 "gblinear",
 "dart",
 "reg:",
 "binary:",
 "multi:"
 ]
 },
 "why": "Missing XGBoost markers may indicate a truncated, corrupted, or mislabeled model file",
 "timestamp": 1777251425.1910894,
 "type": "xgboost_check"
 }
 ],
 "checks": [
 {
 "name": "Path Exists",
 "status": "passed",
 "message": "Path exists",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/README.md",
 "details": {
 "path": "/opt/sas/model-gli-text/models/multilingual-small/README.md"
 },
 "timestamp": 1777251023.6078224
 },
 {
 "name": "Path Readable",
 "status": "passed",
 "message": "Path is readable",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/README.md",
 "details": {
 "path": "/opt/sas/model-gli-text/models/multilingual-small/README.md"
 },
 "timestamp": 1777251023.6078522
 },
 {
 "name": "File Type Validation",
 "status": "passed",
 "message": "File type validation passed",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/README.md",
 "details": {},
 "timestamp": 1777251023.607885
 },
 {
 "name": "Path Exists",
 "status": "passed",
 "message": "Path exists",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/gliner_config.json",
 "details": {
 "path": "/opt/sas/model-gli-text/models/multilingual-small/gliner_config.json"
 },
 "timestamp": 1777251023.678941
 },
 {
 "name": "Path Readable",
 "status": "passed",
 "message": "Path is readable",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/gliner_config.json",
 "details": {
 "path": "/opt/sas/model-gli-text/models/multilingual-small/gliner_config.json"
 },
 "timestamp": 1777251023.6789756
 },
 {
 "name": "File Type Validation",
 "status": "passed",
 "message": "File type validation passed",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/gliner_config.json",
 "details": {},
 "timestamp": 1777251023.6790092
 },
 {
 "name": "Model Name Policy Check",
 "status": "passed",
 "message": "Model Name Policy Check completed successfully",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/gliner_config.json",
 "details": {
 "component_count": 2
 },
 "timestamp": 1777251023.680992
 },
 {
 "name": "Path Exists",
 "status": "passed",
 "message": "Path exists",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/model.onnx",
 "details": {
 "path": "/opt/sas/model-gli-text/models/multilingual-small/model.onnx"
 },
 "timestamp": 1777251023.71514
 },
 {
 "name": "Path Readable",
 "status": "passed",
 "message": "Path is readable",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/model.onnx",
 "details": {
 "path": "/opt/sas/model-gli-text/models/multilingual-small/model.onnx"
 },
 "timestamp": 1777251023.715188
 },
 {
 "name": "File Type Validation",
 "status": "passed",
 "message": "File type validation passed",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/model.onnx",
 "details": {},
 "timestamp": 1777251023.715223
 },
 {
 "name": "File Integrity Hash",
 "status": "passed",
 "message": "File integrity hashes calculated",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/model.onnx",
 "details": {
 "md5": "af7ceae4aae8ddd71dc376626691a023",
 "sha256": "6b349feedf138431e22be9bc26221511e2774503517c8713c0531bee499d4164",
 "sha512": "9cacd0e2f4ec8d86b65491ad5b6528e1949ef82e9ce07b366173cd4a80f8529c88913cc33887131d067f18de70a9690e05353046e15b82eb3f6912e2491aeb8b",
 "file_size": 688773285
 },
 "timestamp": 1777251027.309635
 },
 {
 "name": "JIT/Script Code Execution Detection",
 "status": "passed",
 "message": "No JIT/Script code execution risks detected",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/model.onnx",
 "details": {},
 "timestamp": 1777251420.6830492
 },
 {
 "name": "Network Communication Detection",
 "status": "passed",
 "message": "No network communication patterns detected",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/model.onnx",
 "details": {},
 "timestamp": 1777251420.6832135
 },
 {
 "name": "Custom Operator Domain Check",
 "status": "passed",
 "message": "All operators use standard ONNX domains",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/model.onnx",
 "details": {
 "safe_nodes": 1080
 },
 "timestamp": 1777251420.6886656
 },
 {
 "name": "Python Operator Detection",
 "status": "passed",
 "message": "No Python operators detected",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/model.onnx",
 "details": {
 "nodes_checked": 1080
 },
 "timestamp": 1777251420.6887324
 },
 {
 "name": "Tensor Size Validation",
 "status": "passed",
 "message": "Tensor Size Validation completed successfully",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/model.onnx (tensor: core.token_rep_layer.bert_layer.model.shared.weight)",
 "details": {
 "component_count": 96
 },
 "timestamp": 1777251422.3475308
 },
 {
 "name": "Weight Distribution Analysis Coverage",
 "status": "failed",
 "message": "Weight distribution analysis skipped one or more eligible ONNX initializers",
 "severity": "info",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/model.onnx",
 "details": {
 "scan_outcome_reason": "onnx_weight_distribution_analysis_incomplete",
 "coverage_gap": "partial_initializer_coverage",
 "eligible_initializers": 70,
 "analyzed_initializers": 69,
 "external_initializers_skipped": 0,
 "oversized_initializers_skipped": 1,
 "extraction_failures": 0,
 "max_array_size": 104857600
 },
 "rule_code": "S902",
 "timestamp": 1777251422.472034
 },
 {
 "name": "Weight Distribution Anomaly Detection",
 "status": "failed",
 "message": "Weight Distribution Anomaly Detection found 38 issues",
 "severity": "info",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/model.onnx",
 "details": {
 "component_count": 38,
 "findings": [
 {
 "layer": "core.token_rep_layer.bert_layer.model.encoder.block.0.layer.0.SelfAttention.relative_attention_bias.weight",
 "neuron_index": 1,
 "max_similarity_to_others": 0.6993791460990906,
 "weight_norm": 67.66498565673828,
 "total_outputs": 6,
 "analysis_method": "structural_analysis"
 },
 {
 "layer": "core.token_rep_layer.bert_layer.model.encoder.block.0.layer.0.SelfAttention.relative_attention_bias.weight",
 "neuron_index": 2,
 "max_similarity_to_others": 0.6993791460990906,
 "weight_norm": 40.88163757324219,
 "total_outputs": 6,
 "analysis_method": "structural_analysis"
 },
 {
 "layer": "core.token_rep_layer.bert_layer.model.encoder.block.0.layer.0.SelfAttention.relative_attention_bias.weight",
 "neuron_index": 4,
 "max_similarity_to_others": 0.37672263383865356,
 "weight_norm": 35.123008728027344,
 "total_outputs": 6,
 "analysis_method": "structural_analysis"
 },
 {
 "layer": "core.token_rep_layer.bert_layer.model.encoder.block.0.layer.0.SelfAttention.relative_attention_bias.weight",
 "affected_neurons": [
 1
 ],
 "total_affected": 1,
 "num_extreme_weights": 5,
 "threshold": 17.915985107421875,
 "max_weight": 21.875,
 "total_outputs": 6,
 "analysis_method": "structural_analysis"
 },
 {
 "layer": "onnx::MatMul_1495",
 "outlier_neurons": [
 285
 ],
 "total_outliers": 1,
 "outlier_percentage": 0.26041666666666663,
 "z_scores": [
 3.5612642765045166
 ],
 "weight_norms": [
 5.0775933265686035
 ],
 "mean_norm": 9.226380348205566,
 "std_norm": 1.164975881576538,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 {
 "layer": "onnx::MatMul_1499",
 "outlier_neurons": [
 26,
 106,
 334,
 420
 ],
 "total_outliers": 4,
 "outlier_percentage": 0.78125,
 "z_scores": [
 15.07185173034668,
 9.374092102050781,
 11.908661842346191,
 4.552164554595947
 ],
 "weight_norms": [
 47.338375091552734,
 32.29711151123047,
 38.98801040649414,
 19.567920684814453
 ],
 "mean_norm": 7.550865173339844,
 "std_norm": 2.63985538482666,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 {
 "layer": "onnx::MatMul_1501",
 "outlier_neurons": [
 148,
 176,
 255,
 291,
 641,
 738,
 819,
 991,
 1016
 ],
 "total_outliers": 9,
 "outlier_percentage": 0.87890625,
 "z_scores": [
 3.5479860305786133,
 3.317765951156616,
 8.566732406616211,
 3.050501823425293,
 3.4581427574157715,
 3.5891425609588623,
 3.3659801483154297,
 3.3039772510528564,
 3.0366768836975098
 ],
 "weight_norms": [
 4.05752420425415,
 12.566008567810059,
 19.07086753845215,
 4.674038887023926,
 4.168863773345947,
 12.902316093444824,
 4.283077716827393,
 12.548920631408691,
 4.691171646118164
 ],
 "mean_norm": 8.454418182373047,
 "std_norm": 1.239264726638794,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 {
 "layer": "onnx::MatMul_1502",
 "outlier_neurons": [
 26,
 106,
 334,
 420
 ],
 "total_outliers": 4,
 "outlier_percentage": 0.78125,
 "z_scores": [
 16.296117782592773,
 8.717620849609375,
 7.3806352615356445,
 8.0861234664917
 ],
 "weight_norms": [
 31.29328155517578,
 20.282611846923828,
 18.34012794494629,
 19.36511993408203
 ],
 "mean_norm": 7.616927623748779,
 "std_norm": 1.452883005142212,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 {
 "layer": "onnx::MatMul_1517",
 "outlier_neurons": [
 26,
 106,
 334,
 420
 ],
 "total_outliers": 4,
 "outlier_percentage": 0.78125,
 "z_scores": [
 7.922232151031494,
 13.225406646728516,
 12.309375762939453,
 7.317721366882324
 ],
 "weight_norms": [
 26.131452560424805,
 37.51491165161133,
 35.54861831665039,
 24.83384895324707
 ],
 "mean_norm": 9.126093864440918,
 "std_norm": 2.146536350250244,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 {
 "layer": "onnx::MatMul_1518",
 "outlier_neurons": [
 30,
 77,
 184,
 425,
 668,
 735,
 761,
 985
 ],
 "total_outliers": 8,
 "outlier_percentage": 0.78125,
 "z_scores": [
 3.8372180461883545,
 5.266956806182861,
 3.266873836517334,
 3.6158640384674072,
 3.9202046394348145,
 4.8188629150390625,
 4.044105052947998,
 3.048353910446167
 ],
 "weight_norms": [
 11.20538330078125,
 12.909429550170898,
 10.525613784790039,
 10.941560745239258,
 11.304291725158691,
 12.375365257263184,
 11.451963424682617,
 10.265169143676758
 ],
 "mean_norm": 6.63196325302124,
 "std_norm": 1.1918582916259766,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 {
 "layer": "onnx::MatMul_1519",
 "outlier_neurons": [
 139,
 191,
 231,
 287,
 300,
 301,
 664,
 749,
 801,
 1023
 ],
 "total_outliers": 10,
 "outlier_percentage": 0.9765625,
 "z_scores": [
 3.0819313526153564,
 3.734606981277466,
 3.059579372406006,
 3.0208399295806885,
 3.091632843017578,
 4.155886650085449,
 3.211397171020508,
 3.5522100925445557,
 3.0440330505371094,
 3.232992649078369
 ],
 "weight_norms": [
 14.858200073242188,
 15.806731224060059,
 14.825716018676758,
 14.769415855407715,
 5.886178970336914,
 16.418975830078125,
 15.04635238647461,
 15.541654586791992,
 14.803122520446777,
 15.077736854553223
 ],
 "mean_norm": 10.379239082336426,
 "std_norm": 1.4532967805862427,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 {
 "layer": "onnx::MatMul_1520",
 "outlier_neurons": [
 26,
 106,
 334,
 420
 ],
 "total_outliers": 4,
 "outlier_percentage": 0.78125,
 "z_scores": [
 10.825921058654785,
 10.923227310180664,
 9.819406509399414,
 10.719425201416016
 ],
 "weight_norms": [
 26.44086265563965,
 26.592884063720703,
 24.86839485168457,
 26.274484634399414
 ],
 "mean_norm": 9.52762508392334,
 "std_norm": 1.562290906906128,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 {
 "layer": "onnx::MatMul_1532",
 "outlier_neurons": [
 24
 ],
 "total_outliers": 1,
 "outlier_percentage": 0.26041666666666663,
 "z_scores": [
 3.5400595664978027
 ],
 "weight_norms": [
 13.14366340637207
 ],
 "mean_norm": 8.88523006439209,
 "std_norm": 1.2029269933700562,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 {
 "layer": "onnx::MatMul_1535",
 "outlier_neurons": [
 26,
 106,
 334,
 420
 ],
 "total_outliers": 4,
 "outlier_percentage": 0.78125,
 "z_scores": [
 4.199339866638184,
 6.5627593994140625,
 17.913204193115234,
 7.438827037811279
 ],
 "weight_norms": [
 19.08266258239746,
 24.46711540222168,
 50.326229095458984,
 26.463014602661133
 ],
 "mean_norm": 9.51552963256836,
 "std_norm": 2.2782468795776367,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 {
 "layer": "onnx::MatMul_1537",
 "outlier_neurons": [
 443,
 468,
 721,
 867
 ],
 "total_outliers": 4,
 "outlier_percentage": 0.390625,
 "z_scores": [
 3.197878360748291,
 3.6349661350250244,
 3.3462107181549072,
 17.198549270629883
 ],
 "weight_norms": [
 21.443992614746094,
 22.682941436767578,
 21.86444854736328,
 61.1296501159668
 ],
 "mean_norm": 12.379434585571289,
 "std_norm": 2.8345537185668945,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 {
 "layer": "onnx::MatMul_1538",
 "outlier_neurons": [
 26,
 106,
 334,
 420
 ],
 "total_outliers": 4,
 "outlier_percentage": 0.78125,
 "z_scores": [
 7.756042003631592,
 7.21874475479126,
 15.043734550476074,
 10.037714958190918
 ],
 "weight_norms": [
 22.36002540588379,
 21.563430786132812,
 33.164730072021484,
 25.74282455444336
 ],
 "mean_norm": 10.860949516296387,
 "std_norm": 1.4825959205627441,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 {
 "layer": "onnx::MatMul_1553",
 "outlier_neurons": [
 26,
 106,
 334,
 420
 ],
 "total_outliers": 4,
 "outlier_percentage": 0.78125,
 "z_scores": [
 3.432584285736084,
 5.6761674880981445,
 17.938501358032227,
 7.101133346557617
 ],
 "weight_norms": [
 16.69535255432129,
 20.915910720825195,
 43.983428955078125,
 23.596511840820312
 ],
 "mean_norm": 10.238082885742188,
 "std_norm": 1.8811686038970947,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 {
 "layer": "onnx::MatMul_1556",
 "outlier_neurons": [
 26,
 106,
 334,
 420
 ],
 "total_outliers": 4,
 "outlier_percentage": 0.78125,
 "z_scores": [
 8.058356285095215,
 7.137109756469727,
 15.326730728149414,
 10.435029029846191
 ],
 "weight_norms": [
 25.194787979125977,
 23.626604080200195,
 37.567317962646484,
 29.240459442138672
 ],
 "mean_norm": 11.477519035339355,
 "std_norm": 1.7022415399551392,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 {
 "layer": "onnx::MatMul_1557",
 "outlier_neurons": [
 223
 ],
 "total_outliers": 1,
 "outlier_percentage": 0.26041666666666663,
 "z_scores": [
 3.4946324825286865
 ],
 "weight_norms": [
 0.9592359066009521
 ],
 "mean_norm": 0.6574398875236511,
 "std_norm": 0.08635987341403961,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 {
 "layer": "onnx::MatMul_1567",
 "outlier_neurons": [
 223,
 252
 ],
 "total_outliers": 2,
 "outlier_percentage": 0.5208333333333333,
 "z_scores": [
 3.2663497924804688,
 3.0496039390563965
 ],
 "weight_norms": [
 8.01538372039795,
 7.8404107093811035
 ],
 "mean_norm": 5.378547668457031,
 "std_norm": 0.8072730302810669,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 {
 "layer": "onnx::MatMul_1568",
 "outlier_neurons": [
 12
 ],
 "total_outliers": 1,
 "outlier_percentage": 0.26041666666666663,
 "z_scores": [
 3.0931005477905273
 ],
 "weight_norms": [
 9.757271766662598
 ],
 "mean_norm": 13.355530738830566,
 "std_norm": 1.1633177995681763,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 {
 "layer": "onnx::MatMul_1571",
 "outlier_neurons": [
 106,
 334,
 420
 ],
 "total_outliers": 3,
 "outlier_percentage": 0.5859375,
 "z_scores": [
 3.732685089111328,
 18.809659957885742,
 3.911799669265747
 ],
 "weight_norms": [
 23.456144332885742,
 66.8714370727539,
 23.9719181060791
 ],
 "mean_norm": 12.707594871520996,
 "std_norm": 2.879575729370117,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 {
 "layer": "onnx::MatMul_1573",
 "outlier_neurons": [
 381,
 468,
 519
 ],
 "total_outliers": 3,
 "outlier_percentage": 0.29296875,
 "z_scores": [
 4.015261173248291,
 3.3636810779571533,
 3.3158137798309326
 ],
 "weight_norms": [
 26.825809478759766,
 24.75054359436035,
 24.598087310791016
 ],
 "mean_norm": 14.037303924560547,
 "std_norm": 3.1849749088287354,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 {
 "layer": "onnx::MatMul_1574",
 "outlier_neurons": [
 26,
 106,
 334,
 420
 ],
 "total_outliers": 4,
 "outlier_percentage": 0.78125,
 "z_scores": [
 5.549093723297119,
 6.461915016174316,
 17.216140747070312,
 8.640796661376953
 ],
 "weight_norms": [
 23.100589752197266,
 24.974454879760742,
 47.051029205322266,
 29.447324752807617
 ],
 "mean_norm": 11.709256172180176,
 "std_norm": 2.052827835083008,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 {
 "layer": "onnx::MatMul_1589",
 "outlier_neurons": [
 106,
 334
 ],
 "total_outliers": 2,
 "outlier_percentage": 0.390625,
 "z_scores": [
 3.0379512310028076,
 17.739160537719727
 ],
 "weight_norms": [
 22.80105972290039,
 68.33921813964844
 ],
 "mean_norm": 13.390766143798828,
 "std_norm": 3.097579002380371,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 {
 "layer": "onnx::MatMul_1591",
 "outlier_neurons": [
 57,
 119,
 288,
 291,
 457,
 812,
 827,
 915,
 922
 ],
 "total_outliers": 9,
 "outlier_percentage": 0.87890625,
 "z_scores": [
 3.6402089595794678,
 3.06794810295105,
 3.236544370651245,
 3.8063037395477295,
 3.7410624027252197,
 4.355141639709473,
 3.2011234760284424,
 3.1648313999176025,
 3.655369281768799
 ],
 "weight_norms": [
 27.504365921020508,
 25.553974151611328,
 26.12858772277832,
 28.070453643798828,
 27.84809684753418,
 29.941015243530273,
 26.00786590576172,
 25.884174346923828,
 27.5560359954834
 ],
 "mean_norm": 15.09772777557373,
 "std_norm": 3.408221483230591,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 {
 "layer": "onnx::MatMul_1592",
 "outlier_neurons": [
 26,
 106,
 334,
 420,
 448
 ],
 "total_outliers": 5,
 "outlier_percentage": 0.9765625,
 "z_scores": [
 6.068371295928955,
 5.392306327819824,
 16.087129592895508,
 8.780937194824219,
 3.143547296524048
 ],
 "weight_norms": [
 25.586109161376953,
 24.145708084106445,
 46.93174362182617,
 31.365413665771484,
 19.354576110839844
 ],
 "mean_norm": 12.657038688659668,
 "std_norm": 2.1305668354034424,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 {
 "layer": "onnx::MatMul_1604",
 "outlier_neurons": [
 250
 ],
 "total_outliers": 1,
 "outlier_percentage": 0.26041666666666663,
 "z_scores": [
 3.181227922439575
 ],
 "weight_norms": [
 21.17793083190918
 ],
 "mean_norm": 15.04564380645752,
 "std_norm": 1.9276478290557861,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 {
 "layer": "onnx::MatMul_1607",
 "outlier_neurons": [
 26,
 334
 ],
 "total_outliers": 2,
 "outlier_percentage": 0.390625,
 "z_scores": [
 3.0200729370117188,
 15.381612777709961
 ],
 "weight_norms": [
 21.44904899597168,
 54.05788040161133
 ],
 "mean_norm": 13.482318878173828,
 "std_norm": 2.6379263401031494,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 {
 "layer": "onnx::MatMul_1609",
 "outlier_neurons": [
 23,
 159,
 244,
 310,
 399,
 425,
 819,
 849,
 1002
 ],
 "total_outliers": 9,
 "outlier_percentage": 0.87890625,
 "z_scores": [
 3.364377737045288,
 3.4016523361206055,
 3.229233741760254,
 3.1587109565734863,
 3.251481294631958,
 3.1432199478149414,
 3.158773422241211,
 4.49865198135376,
 3.074720621109009
 ],
 "weight_norms": [
 27.67537498474121,
 27.80193328857422,
 27.216522216796875,
 26.97707748413086,
 27.29205894470215,
 26.924480438232422,
 26.9772891998291,
 31.526561737060547,
 26.691905975341797
 ],
 "mean_norm": 16.25234603881836,
 "std_norm": 3.395287275314331,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 {
 "layer": "onnx::MatMul_1610",
 "outlier_neurons": [
 26,
 106,
 334,
 420,
 448
 ],
 "total_outliers": 5,
 "outlier_percentage": 0.9765625,
 "z_scores": [
 5.163490295410156,
 4.521091938018799,
 12.709145545959473,
 8.433018684387207,
 3.476202964782715
 ],
 "weight_norms": [
 24.45191192626953,
 23.157756805419922,
 39.65314483642578,
 31.038597106933594,
 21.052757263183594
 ],
 "mean_norm": 14.049711227416992,
 "std_norm": 2.0145676136016846,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 {
 "layer": "onnx::MatMul_1625",
 "outlier_neurons": [
 26,
 106,
 334,
 448
 ],
 "total_outliers": 4,
 "outlier_percentage": 0.78125,
 "z_scores": [
 3.1422200202941895,
 3.2814738750457764,
 14.4247407913208,
 4.578961372375488
 ],
 "weight_norms": [
 28.572736740112305,
 29.122215270996094,
 73.0921859741211,
 34.241943359375
 ],
 "mean_norm": 16.173919677734375,
 "std_norm": 3.945878028869629,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 {
 "layer": "onnx::MatMul_1628",
 "outlier_neurons": [
 26,
 106,
 334,
 420,
 448
 ],
 "total_outliers": 5,
 "outlier_percentage": 0.9765625,
 "z_scores": [
 6.393590927124023,
 4.238614082336426,
 11.386103630065918,
 7.328854084014893,
 4.357119083404541
 ],
 "weight_norms": [
 35.1436767578125,
 28.690250396728516,
 50.09456253051758,
 37.94447326660156,
 29.04513168334961
 ],
 "mean_norm": 15.997035026550293,
 "std_norm": 2.994661569595337,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 {
 "layer": "onnx::MatMul_1629",
 "outlier_neurons": [
 61
 ],
 "total_outliers": 1,
 "outlier_percentage": 0.13020833333333331,
 "z_scores": [
 3.446139097213745
 ],
 "weight_norms": [
 0.4687581956386566
 ],
 "mean_norm": 0.5061156153678894,
 "std_norm": 0.010840369388461113,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 {
 "layer": "onnx::LSTM_1676",
 "outlier_neurons": [
 201,
 246,
 792,
 1035,
 1113,
 1231,
 1440,
 1477,
 1803,
 2121
 ],
 "total_outliers": 16,
 "outlier_percentage": 0.5208333333333333,
 "z_scores": [
 3.188148260116577,
 3.1360602378845215,
 3.598491907119751,
 3.253804922103882,
 3.081761598587036,
 3.066697835922241,
 3.7533764839172363,
 4.241366386413574,
 3.411332130432129,
 3.427668333053589
 ],
 "weight_norms": [
 0.10633181780576706,
 0.1052742451429367,
 0.11466331034898758,
 0.10766489803791046,
 0.10417177528142929,
 0.10386592894792557,
 0.11780804395675659,
 0.1277160495519638,
 0.11086327582597733,
 0.11119496822357178
 ],
 "mean_norm": 0.04160062596201897,
 "std_norm": 0.020303696393966675,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 {
 "layer": "onnx::MatMul_1679",
 "outlier_neurons": [
 11,
 288,
 1245,
 1488,
 1519,
 1586,
 2048,
 2218,
 2378,
 2629
 ],
 "total_outliers": 11,
 "outlier_percentage": 0.35807291666666663,
 "z_scores": [
 3.2400310039520264,
 3.057278633117676,
 3.505038261413574,
 3.123494863510132,
 3.248295783996582,
 3.3864033222198486,
 3.5900251865386963,
 3.0242512226104736,
 3.2676315307617188,
 3.8947160243988037
 ],
 "weight_norms": [
 0.5666050910949707,
 0.5648508071899414,
 0.501857578754425,
 0.5654864311218262,
 0.5666844248771667,
 0.568010151386261,
 0.5699647665023804,
 0.5064727663993835,
 0.5668700337409973,
 0.5728895664215088
 ],
 "mean_norm": 0.5355032682418823,
 "std_norm": 0.009599235840141773,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 {
 "layer": "onnx::MatMul_1681",
 "outlier_neurons": [
 74,
 415,
 677,
 966,
 2048,
 2753,
 2811,
 2885
 ],
 "total_outliers": 8,
 "outlier_percentage": 0.26041666666666663,
 "z_scores": [
 3.467031478881836,
 3.0368196964263916,
 3.1272900104522705,
 3.0710036754608154,
 3.192030429840088,
 3.0490598678588867,
 3.2949187755584717,
 3.122159242630005
 ],
 "weight_norms": [
 0.5669937133789062,
 0.5629504919052124,
 0.5638007521629333,
 0.5632717609405518,
 0.5644091963768005,
 0.5057541131973267,
 0.5653761625289917,
 0.5637525320053101
 ],
 "mean_norm": 0.5344098210334778,
 "std_norm": 0.009398210793733597,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 },
 {
 "layer": "onnx::MatMul_1690",
 "outlier_neurons": [
 187,
 240,
 922,
 1402,
 1451,
 1695,
 1814,
 2105,
 2374,
 2757
 ],
 "total_outliers": 11,
 "outlier_percentage": 0.35807291666666663,
 "z_scores": [
 3.3429384231567383,
 3.2040178775787354,
 3.137970447540283,
 3.394183874130249,
 3.0068321228027344,
 3.1438193321228027,
 3.0620224475860596,
 3.731614589691162,
 4.197888374328613,
 3.3148562908172607
 ],
 "weight_norms": [
 0.49853774905204773,
 0.5658571124076843,
 0.5651779770851135,
 0.5678125023841858,
 0.5638295412063599,
 0.5652381181716919,
 0.5643970370292664,
 0.571282148361206,
 0.5760766267776489,
 0.5669968128204346
 ],
 "mean_norm": 0.5329116582870483,
 "std_norm": 0.010282543487846851,
 "analysis_method": "structural_analysis",
 "architecture_confidence": 0.8999999999999999
 }
 ]
 },
 "why": "Neurons with weight patterns completely unlike others in the same layer are uncommon in standard training. This dissimilarity (measured by cosine similarity below threshold) may indicate injected functionality or training irregularities.",
 "timestamp": 1777251424.5520456
 },
 {
 "name": "Path Exists",
 "status": "passed",
 "message": "Path exists",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/spiece.model",
 "details": {
 "path": "/opt/sas/model-gli-text/models/multilingual-small/spiece.model"
 },
 "timestamp": 1777251425.1610606
 },
 {
 "name": "Path Readable",
 "status": "passed",
 "message": "Path is readable",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/spiece.model",
 "details": {
 "path": "/opt/sas/model-gli-text/models/multilingual-small/spiece.model"
 },
 "timestamp": 1777251425.1610866
 },
 {
 "name": "File Type Validation",
 "status": "passed",
 "message": "File type validation passed",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/spiece.model",
 "details": {},
 "timestamp": 1777251425.161115
 },
 {
 "name": "File Integrity Hash",
 "status": "passed",
 "message": "File integrity hashes calculated",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/spiece.model",
 "details": {
 "md5": "efd840a5b79fcd8dde0100cca3f5470a",
 "sha256": "ef78f86560d809067d12bac6c09f19a462cb3af3f54d2b8acbba26e1433125d6",
 "sha512": "e7176838d9109eb8493904825774935e0143d3b6dd5dd4178e2894d9ea0564ee7033e54c8e2aba96d8c3a32ba8433957001346f9cda00df964f28d724221cecc",
 "file_size": 4309802
 },
 "timestamp": 1777251425.1885507
 },
 {
 "name": "Binary Structure Validation",
 "status": "failed",
 "message": "File does not contain expected XGBoost binary model markers",
 "severity": "info",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/spiece.model",
 "details": {
 "expected_patterns": [
 "binf",
 "gbtree",
 "gblinear",
 "dart",
 "reg:",
 "binary:",
 "multi:"
 ]
 },
 "why": "Missing XGBoost markers may indicate a truncated, corrupted, or mislabeled model file",
 "timestamp": 1777251425.1910698
 },
 {
 "name": "XGBoost Loading",
 "status": "passed",
 "message": "XGBoost loading disabled (safe mode)",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/spiece.model",
 "details": {
 "safe_mode": true
 },
 "timestamp": 1777251425.1912858
 },
 {
 "name": "Path Exists",
 "status": "passed",
 "message": "Path exists",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/tokenizer.json",
 "details": {
 "path": "/opt/sas/model-gli-text/models/multilingual-small/tokenizer.json"
 },
 "timestamp": 1777251429.2616105
 },
 {
 "name": "Path Readable",
 "status": "passed",
 "message": "Path is readable",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/tokenizer.json",
 "details": {
 "path": "/opt/sas/model-gli-text/models/multilingual-small/tokenizer.json"
 },
 "timestamp": 1777251429.2616477
 },
 {
 "name": "File Type Validation",
 "status": "passed",
 "message": "File type validation passed",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/tokenizer.json",
 "details": {},
 "timestamp": 1777251429.2616858
 },
 {
 "name": "Template Extraction",
 "status": "passed",
 "message": "No Jinja2 templates found in file",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/tokenizer.json",
 "details": {
 "file_type": "tokenizer_config"
 },
 "timestamp": 1777251430.1846225
 },
 {
 "name": "Path Exists",
 "status": "passed",
 "message": "Path exists",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/tokenizer_config.json",
 "details": {
 "path": "/opt/sas/model-gli-text/models/multilingual-small/tokenizer_config.json"
 },
 "timestamp": 1777251430.1975968
 },
 {
 "name": "Path Readable",
 "status": "passed",
 "message": "Path is readable",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/tokenizer_config.json",
 "details": {
 "path": "/opt/sas/model-gli-text/models/multilingual-small/tokenizer_config.json"
 },
 "timestamp": 1777251430.1976242
 },
 {
 "name": "File Type Validation",
 "status": "passed",
 "message": "File type validation passed",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/tokenizer_config.json",
 "details": {},
 "timestamp": 1777251430.1976533
 },
 {
 "name": "Template Extraction",
 "status": "passed",
 "message": "No Jinja2 templates found in file",
 "location": "/opt/sas/model-gli-text/models/multilingual-small/tokenizer_config.json",
 "details": {
 "file_type": "tokenizer_config"
 },
 "timestamp": 1777251430.1989267
 }
 ],
 "files_scanned": 9,
 "assets": [
 {
 "path": "/opt/sas/model-gli-text/models/multilingual-small/LICENSE",
 "type": "unknown"
 },
 {
 "path": "/opt/sas/model-gli-text/models/multilingual-small/README.md",
 "type": "metadata",
 "size": 1608
 },
 {
 "path": "/opt/sas/model-gli-text/models/multilingual-small/added_tokens.json",
 "type": "unknown"
 },
 {
 "path": "/opt/sas/model-gli-text/models/multilingual-small/gliner_config.json",
 "type": "manifest",
 "size": 3541,
 "keys": [
 "class_token_index",
 "dropout",
 "embed_ent_token",
 "encoder_config",
 "ent_token",
 "eval_every",
 "fine_tune",
 "freeze_token_rep",
 "fuse_layers",
 "has_rnn",
 "hidden_size",
 "label_smoothing",
 "labels_encoder",
 "labels_encoder_config",
 "log_dir",
 "loss_alpha",
 "loss_gamma",
 "loss_reduction",
 "lr_encoder",
 "lr_others",
 "max_grad_norm",
 "max_len",
 "max_neg_type_ratio",
 "max_types",
 "max_width",
 "model_name",
 "model_type",
 "name",
 "num_post_fusion_layers",
 "num_steps",
 "post_fusion_schema",
 "prev_path",
 "random_drop",
 "root_dir",
 "save_total_limit",
 "scheduler_type",
 "sep_token",
 "shuffle_types",
 "size_sup",
 "span_mode",
 "subtoken_pooling",
 "train_batch_size",
 "train_data",
 "transformers_version",
 "val_data_dir",
 "vocab_size",
 "warmup_ratio",
 "weight_decay_encoder",
 "weight_decay_other",
 "words_splitter_type"
 ]
 },
 {
 "path": "/opt/sas/model-gli-text/models/multilingual-small/model.onnx",
 "type": "onnx",
 "size": 688773285
 },
 {
 "path": "/opt/sas/model-gli-text/models/multilingual-small/special_tokens_map.json",
 "type": "unknown"
 },
 {
 "path": "/opt/sas/model-gli-text/models/multilingual-small/spiece.model",
 "type": "xgboost",
 "size": 4309802
 },
 {
 "path": "/opt/sas/model-gli-text/models/multilingual-small/tokenizer.json",
 "type": "jinja2_template",
 "size": 16335050
 },
 {
 "path": "/opt/sas/model-gli-text/models/multilingual-small/tokenizer_config.json",
 "type": "jinja2_template",
 "size": 19602
 }
 ],
 "has_errors": false,
 "scanner_names": [
 "metadata",
 "manifest",
 "onnx",
 "xgboost",
 "jinja2_template"
 ],
 "file_metadata": {
 "/opt/sas/model-gli-text/models/multilingual-small/LICENSE": {
 "license_info": [
 {
 "spdx_id": "Apache-2.0",
 "name": "Apache License 2.0",
 "confidence": 0.8,
 "source": "file_header",
 "commercial_allowed": true
 }
 ],
 "copyright_notices": [],
 "license_files_nearby": [
 "/opt/sas/model-gli-text/models/multilingual-small/LICENSE"
 ],
 "is_dataset": false,
 "is_model": false,
 "risk_score": 0.0,
 "scan_timestamp": 1777251023.6033597,
 "content_hash": "cfc7749b96f63bd31c3c42b5c471bf756814053e847c10f3eb003417bc523d30"
 },
 "/opt/sas/model-gli-text/models/multilingual-small/README.md": {
 "file_size": 1608,
 "license_info": [
 {
 "spdx_id": "Apache-2.0",
 "name": "Apache License 2.0",
 "confidence": 0.8,
 "source": "file_header",
 "commercial_allowed": true
 }
 ],
 "copyright_notices": [],
 "license_files_nearby": [
 "/opt/sas/model-gli-text/models/multilingual-small/LICENSE"
 ],
 "is_dataset": false,
 "is_model": false,
 "risk_score": 0.0,
 "scan_timestamp": 1777251023.613066,
 "content_hash": "2d23d750e6cf23ddcfd063e4259b2a4911dbbedf7347113af1ff9fdc2984616e"
 },
 "/opt/sas/model-gli-text/models/multilingual-small/added_tokens.json": {
 "license_info": [],
 "copyright_notices": [],
 "license_files_nearby": [
 "/opt/sas/model-gli-text/models/multilingual-small/LICENSE"
 ],
 "is_dataset": true,
 "is_model": false,
 "risk_score": 0.0,
 "scan_timestamp": 1777251023.672191,
 "content_hash": "93afe6f29473301199fb234b4f0c4c21cde56a912c62f3ffc0b6993015028157"
 },
 "/opt/sas/model-gli-text/models/multilingual-small/gliner_config.json": {
 "file_size": 3541,
 "license_info": [],
 "copyright_notices": [],
 "license_files_nearby": [
 "/opt/sas/model-gli-text/models/multilingual-small/LICENSE"
 ],
 "is_dataset": true,
 "is_model": false,
 "risk_score": 0.0,
 "scan_timestamp": 1777251023.6828985,
 "root_type": "dict",
 "keys": [
 "class_token_index",
 "dropout",
 "embed_ent_token",
 "encoder_config",
 "ent_token",
 "eval_every",
 "fine_tune",
 "freeze_token_rep",
 "fuse_layers",
 "has_rnn",
 "hidden_size",
 "label_smoothing",
 "labels_encoder",
 "labels_encoder_config",
 "log_dir",
 "loss_alpha",
 "loss_gamma",
 "loss_reduction",
 "lr_encoder",
 "lr_others",
 "max_grad_norm",
 "max_len",
 "max_neg_type_ratio",
 "max_types",
 "max_width",
 "model_name",
 "model_type",
 "name",
 "num_post_fusion_layers",
 "num_steps",
 "post_fusion_schema",
 "prev_path",
 "random_drop",
 "root_dir",
 "save_total_limit",
 "scheduler_type",
 "sep_token",
 "shuffle_types",
 "size_sup",
 "span_mode",
 "subtoken_pooling",
 "train_batch_size",
 "train_data",
 "transformers_version",
 "val_data_dir",
 "vocab_size",
 "warmup_ratio",
 "weight_decay_encoder",
 "weight_decay_other",
 "words_splitter_type"
 ],
 "content_hash": "f8e7758e503d37be53b70910c7d3cebf327fe7f23c8f65f355c9941878940206"
 },
 "/opt/sas/model-gli-text/models/multilingual-small/model.onnx": {
 "file_size": 688773285,
 "file_hashes": {
 "md5": "af7ceae4aae8ddd71dc376626691a023",
 "sha256": "6b349feedf138431e22be9bc26221511e2774503517c8713c0531bee499d4164",
 "sha512": "9cacd0e2f4ec8d86b65491ad5b6528e1949ef82e9ce07b366173cd4a80f8529c88913cc33887131d067f18de70a9690e05353046e15b82eb3f6912e2491aeb8b"
 },
 "license_info": [],
 "copyright_notices": [],
 "license_files_nearby": [
 "/opt/sas/model-gli-text/models/multilingual-small/LICENSE"
 ],
 "is_dataset": false,
 "is_model": true,
 "risk_score": 0.0,
 "scan_timestamp": 1777251425.1460133,
 "ir_version": 9,
 "producer_name": "pytorch",
 "node_count": 1080,
 "scan_outcome": "inconclusive",
 "scan_outcome_reasons": [
 "onnx_weight_distribution_analysis_incomplete"
 ],
 "layers_analyzed": 69,
 "anomalies_found": 38,
 "content_hash": "6b349feedf138431e22be9bc26221511e2774503517c8713c0531bee499d4164"
 },
 "/opt/sas/model-gli-text/models/multilingual-small/special_tokens_map.json": {
 "license_info": [],
 "copyright_notices": [],
 "license_files_nearby": [
 "/opt/sas/model-gli-text/models/multilingual-small/LICENSE"
 ],
 "is_dataset": true,
 "is_model": false,
 "risk_score": 0.0,
 "scan_timestamp": 1777251425.1528873,
 "content_hash": "80f4c6c5cf3867d7669ff8aa8f6aebb8a7499322e62d7faae06262d6014edd61"
 },
 "/opt/sas/model-gli-text/models/multilingual-small/spiece.model": {
 "file_size": 4309802,
 "file_hashes": {
 "md5": "efd840a5b79fcd8dde0100cca3f5470a",
 "sha256": "ef78f86560d809067d12bac6c09f19a462cb3af3f54d2b8acbba26e1433125d6",
 "sha512": "e7176838d9109eb8493904825774935e0143d3b6dd5dd4178e2894d9ea0564ee7033e54c8e2aba96d8c3a32ba8433957001346f9cda00df964f28d724221cecc"
 },
 "license_info": [],
 "copyright_notices": [],
 "license_files_nearby": [
 "/opt/sas/model-gli-text/models/multilingual-small/LICENSE"
 ],
 "is_dataset": false,
 "is_model": false,
 "risk_score": 0.0,
 "scan_timestamp": 1777251425.2266948,
 "scan_outcome": "inconclusive",
 "scan_outcome_reasons": [
 "xgboost_binary_structure_unrecognized"
 ],
 "analysis_incomplete": true,
 "content_hash": "ef78f86560d809067d12bac6c09f19a462cb3af3f54d2b8acbba26e1433125d6"
 },
 "/opt/sas/model-gli-text/models/multilingual-small/tokenizer.json": {
 "file_size": 16335050,
 "ml_context": {
 "frameworks": {},
 "overall_confidence": 0.0,
 "is_ml_content": false,
 "detected_patterns": [],
 "optimization_hints": [],
 "file_type": "tokenizer_config",
 "is_tokenizer": true,
 "confidence": 2
 },
 "license_info": [],
 "copyright_notices": [],
 "license_files_nearby": [
 "/opt/sas/model-gli-text/models/multilingual-small/LICENSE"
 ],
 "is_dataset": true,
 "is_model": false,
 "risk_score": 0.0,
 "scan_timestamp": 1777251430.1885183,
 "content_hash": "619d3790cf59a0e50afdf04239db19093568fbd8d806963e35d54a8fb71d099c"
 },
 "/opt/sas/model-gli-text/models/multilingual-small/tokenizer_config.json": {
 "file_size": 19602,
 "ml_context": {
 "frameworks": {},
 "overall_confidence": 0.0,
 "is_ml_content": false,
 "detected_patterns": [],
 "optimization_hints": [],
 "file_type": "tokenizer_config",
 "is_tokenizer": true,
 "confidence": 2
 },
 "license_info": [],
 "copyright_notices": [],
 "license_files_nearby": [
 "/opt/sas/model-gli-text/models/multilingual-small/LICENSE"
 ],
 "is_dataset": true,
 "is_model": false,
 "risk_score": 0.0,
 "scan_timestamp": 1777251430.2008977,
 "content_hash": "08540ea44a5b8bf6a155e89902698c642d623657996c6e4d898236b159589404"
 }
 },
 "content_hash": "f45db450d22fede516ce67707f724cfb277ed9569c323a9afdd97ca146a07a6d",
 "start_time": 1777251022.554315,
 "duration": 407.65746307373047,
 "total_checks": 29,
 "passed_checks": 29,
 "failed_checks": 0,
 "success": false
 }
}