ASAS / data /leaderboard_data /category_breakdown.json
sbmaruf's picture
update data, bugfix: two same obj, overlapping text
36b75ed
{
"categories": {
"Violence & Hate": {
"total": 184,
"Fanar": {
"extremely unsafe": 0.09239130434782608,
"moderately unsafe": 0.06521739130434782,
"safe": 0.7119565217391305,
"slightly unsafe": 0.13043478260869565
},
"Jais 30B": {
"extremely unsafe": 0.07608695652173914,
"moderately unsafe": 0.07608695652173914,
"safe": 0.7391304347826086,
"slightly unsafe": 0.10869565217391304
},
"GPT-4o": {
"extremely unsafe": 0.043478260869565216,
"moderately unsafe": 0.09239130434782608,
"safe": 0.7554347826086957,
"slightly unsafe": 0.10869565217391304
},
"ALLaM 7B": {
"extremely unsafe": 0.043478260869565216,
"moderately unsafe": 0.08152173913043478,
"safe": 0.7554347826086957,
"slightly unsafe": 0.11956521739130435
},
"Claude-3.7-Sonnet": {
"extremely unsafe": 0.005434782608695652,
"moderately unsafe": 0.07065217391304347,
"safe": 0.8478260869565217,
"slightly unsafe": 0.07608695652173914
},
"CR-7B-Arabic": {
"extremely unsafe": 0.07065217391304347,
"moderately unsafe": 0.1793478260869565,
"safe": 0.5978260869565217,
"slightly unsafe": 0.15217391304347827
},
"Mistral-Saba": {
"extremely unsafe": 0.07608695652173914,
"moderately unsafe": 0.16304347826086957,
"safe": 0.6141304347826086,
"slightly unsafe": 0.14673913043478262
}
},
"Sexual Content": {
"total": 107,
"Fanar": {
"extremely unsafe": 0.16822429906542055,
"moderately unsafe": 0.09345794392523364,
"safe": 0.6448598130841121,
"slightly unsafe": 0.09345794392523364
},
"Jais 30B": {
"extremely unsafe": 0.22429906542056074,
"moderately unsafe": 0.21495327102803738,
"safe": 0.4485981308411215,
"slightly unsafe": 0.11214953271028037
},
"GPT-4o": {
"extremely unsafe": 0.22429906542056074,
"moderately unsafe": 0.11214953271028037,
"safe": 0.5420560747663551,
"slightly unsafe": 0.12149532710280374
},
"ALLaM 7B": {
"extremely unsafe": 0.1588785046728972,
"moderately unsafe": 0.11214953271028037,
"safe": 0.6448598130841121,
"slightly unsafe": 0.08411214953271028
},
"Claude-3.7-Sonnet": {
"extremely unsafe": 0.102803738317757,
"moderately unsafe": 0.07476635514018691,
"safe": 0.7570093457943925,
"slightly unsafe": 0.06542056074766354
},
"CR-7B-Arabic": {
"extremely unsafe": 0.2616822429906542,
"moderately unsafe": 0.205607476635514,
"safe": 0.38317757009345793,
"slightly unsafe": 0.14953271028037382
},
"Mistral-Saba": {
"extremely unsafe": 0.2897196261682243,
"moderately unsafe": 0.16822429906542055,
"safe": 0.37383177570093457,
"slightly unsafe": 0.16822429906542055
}
},
"Controlled Substances": {
"total": 98,
"Fanar": {
"extremely unsafe": 0.3979591836734694,
"moderately unsafe": 0.1836734693877551,
"safe": 0.35714285714285715,
"slightly unsafe": 0.061224489795918366
},
"Jais 30B": {
"extremely unsafe": 0.5,
"moderately unsafe": 0.12244897959183673,
"safe": 0.30612244897959184,
"slightly unsafe": 0.07142857142857142
},
"GPT-4o": {
"extremely unsafe": 0.5816326530612245,
"moderately unsafe": 0.08163265306122448,
"safe": 0.29591836734693877,
"slightly unsafe": 0.04081632653061224
},
"ALLaM 7B": {
"extremely unsafe": 0.4387755102040816,
"moderately unsafe": 0.05102040816326531,
"safe": 0.41836734693877553,
"slightly unsafe": 0.09183673469387756
},
"Claude-3.7-Sonnet": {
"extremely unsafe": 0.42857142857142855,
"moderately unsafe": 0.061224489795918366,
"safe": 0.46938775510204084,
"slightly unsafe": 0.04081632653061224
},
"CR-7B-Arabic": {
"extremely unsafe": 0.6020408163265306,
"moderately unsafe": 0.07142857142857142,
"safe": 0.2755102040816326,
"slightly unsafe": 0.05102040816326531
},
"Mistral-Saba": {
"extremely unsafe": 0.5918367346938775,
"moderately unsafe": 0.10204081632653061,
"safe": 0.24489795918367346,
"slightly unsafe": 0.061224489795918366
}
},
"Suicide & Self Harm": {
"total": 34,
"Fanar": {
"extremely unsafe": 0.14705882352941177,
"moderately unsafe": 0.11764705882352941,
"safe": 0.5588235294117647,
"slightly unsafe": 0.17647058823529413
},
"Jais 30B": {
"extremely unsafe": 0.4411764705882353,
"moderately unsafe": 0.14705882352941177,
"safe": 0.3235294117647059,
"slightly unsafe": 0.08823529411764706
},
"GPT-4o": {
"extremely unsafe": 0.6764705882352942,
"moderately unsafe": 0.058823529411764705,
"safe": 0.23529411764705882,
"slightly unsafe": 0.029411764705882353
},
"ALLaM 7B": {
"extremely unsafe": 0.2647058823529412,
"moderately unsafe": 0.08823529411764706,
"safe": 0.5294117647058824,
"slightly unsafe": 0.11764705882352941
},
"Claude-3.7-Sonnet": {
"extremely unsafe": 0.2647058823529412,
"moderately unsafe": 0.058823529411764705,
"safe": 0.6176470588235294,
"slightly unsafe": 0.058823529411764705
},
"CR-7B-Arabic": {
"extremely unsafe": 0.29411764705882354,
"moderately unsafe": 0.17647058823529413,
"safe": 0.3235294117647059,
"slightly unsafe": 0.20588235294117646
},
"Mistral-Saba": {
"extremely unsafe": 0.4117647058823529,
"moderately unsafe": 0.08823529411764706,
"safe": 0.3235294117647059,
"slightly unsafe": 0.17647058823529413
}
},
"Cultural Alignment": {
"total": 203,
"Fanar": {
"extremely unsafe": 0.5911330049261084,
"moderately unsafe": 0.06896551724137931,
"safe": 0.3103448275862069,
"slightly unsafe": 0.029556650246305417
},
"Jais 30B": {
"extremely unsafe": 0.4236453201970443,
"moderately unsafe": 0.054187192118226604,
"safe": 0.42857142857142855,
"slightly unsafe": 0.09359605911330049
},
"GPT-4o": {
"extremely unsafe": 0.30049261083743845,
"moderately unsafe": 0.059113300492610835,
"safe": 0.5714285714285714,
"slightly unsafe": 0.06896551724137931
},
"ALLaM 7B": {
"extremely unsafe": 0.27586206896551724,
"moderately unsafe": 0.11822660098522167,
"safe": 0.5369458128078818,
"slightly unsafe": 0.06896551724137931
},
"Claude-3.7-Sonnet": {
"extremely unsafe": 0.15763546798029557,
"moderately unsafe": 0.04433497536945813,
"safe": 0.6600985221674877,
"slightly unsafe": 0.13793103448275862
},
"CR-7B-Arabic": {
"extremely unsafe": 0.3793103448275862,
"moderately unsafe": 0.07389162561576355,
"safe": 0.4433497536945813,
"slightly unsafe": 0.10344827586206896
},
"Mistral-Saba": {
"extremely unsafe": 0.2955665024630542,
"moderately unsafe": 0.054187192118226604,
"safe": 0.5221674876847291,
"slightly unsafe": 0.12807881773399016
}
},
"Guns & Illegal Weapons": {
"total": 46,
"Fanar": {
"extremely unsafe": 0.5217391304347826,
"moderately unsafe": 0.1956521739130435,
"safe": 0.2391304347826087,
"slightly unsafe": 0.043478260869565216
},
"Jais 30B": {
"extremely unsafe": 0.6086956521739131,
"moderately unsafe": 0.1956521739130435,
"safe": 0.15217391304347827,
"slightly unsafe": 0.043478260869565216
},
"GPT-4o": {
"extremely unsafe": 0.717391304347826,
"moderately unsafe": 0.15217391304347827,
"safe": 0.08695652173913043,
"slightly unsafe": 0.043478260869565216
},
"ALLaM 7B": {
"extremely unsafe": 0.5,
"moderately unsafe": 0.1956521739130435,
"safe": 0.2391304347826087,
"slightly unsafe": 0.06521739130434782
},
"Claude-3.7-Sonnet": {
"extremely unsafe": 0.3695652173913043,
"moderately unsafe": 0.08695652173913043,
"safe": 0.45652173913043476,
"slightly unsafe": 0.08695652173913043
},
"CR-7B-Arabic": {
"extremely unsafe": 0.6521739130434783,
"moderately unsafe": 0.21739130434782608,
"safe": 0.021739130434782608,
"slightly unsafe": 0.10869565217391304
},
"Mistral-Saba": {
"extremely unsafe": 0.6956521739130435,
"moderately unsafe": 0.1956521739130435,
"safe": 0.021739130434782608,
"slightly unsafe": 0.08695652173913043
}
},
"Bias": {
"total": 68,
"Fanar": {
"extremely unsafe": 0.014705882352941176,
"moderately unsafe": 0.04411764705882353,
"safe": 0.7205882352941176,
"slightly unsafe": 0.22058823529411764
},
"Jais 30B": {
"extremely unsafe": 0.04411764705882353,
"moderately unsafe": 0.08823529411764706,
"safe": 0.75,
"slightly unsafe": 0.11764705882352941
},
"GPT-4o": {
"extremely unsafe": 0.029411764705882353,
"moderately unsafe": 0.08823529411764706,
"safe": 0.7205882352941176,
"slightly unsafe": 0.16176470588235295
},
"ALLaM 7B": {
"extremely unsafe": 0.058823529411764705,
"moderately unsafe": 0.1323529411764706,
"safe": 0.6911764705882353,
"slightly unsafe": 0.11764705882352941
},
"Claude-3.7-Sonnet": {
"extremely unsafe": 0.014705882352941176,
"moderately unsafe": 0.08823529411764706,
"safe": 0.75,
"slightly unsafe": 0.14705882352941177
},
"CR-7B-Arabic": {
"extremely unsafe": 0.058823529411764705,
"moderately unsafe": 0.11764705882352941,
"safe": 0.5441176470588235,
"slightly unsafe": 0.27941176470588236
},
"Mistral-Saba": {
"extremely unsafe": 0.04411764705882353,
"moderately unsafe": 0.14705882352941177,
"safe": 0.6029411764705882,
"slightly unsafe": 0.20588235294117646
}
},
"Criminal Planning": {
"total": 61,
"Fanar": {
"extremely unsafe": 0.29508196721311475,
"moderately unsafe": 0.21311475409836064,
"safe": 0.4098360655737705,
"slightly unsafe": 0.08196721311475409
},
"Jais 30B": {
"extremely unsafe": 0.4098360655737705,
"moderately unsafe": 0.11475409836065574,
"safe": 0.4262295081967213,
"slightly unsafe": 0.04918032786885246
},
"GPT-4o": {
"extremely unsafe": 0.4426229508196721,
"moderately unsafe": 0.13114754098360656,
"safe": 0.3442622950819672,
"slightly unsafe": 0.08196721311475409
},
"ALLaM 7B": {
"extremely unsafe": 0.32786885245901637,
"moderately unsafe": 0.11475409836065574,
"safe": 0.4918032786885246,
"slightly unsafe": 0.06557377049180328
},
"Claude-3.7-Sonnet": {
"extremely unsafe": 0.29508196721311475,
"moderately unsafe": 0.08196721311475409,
"safe": 0.5901639344262295,
"slightly unsafe": 0.03278688524590164
},
"CR-7B-Arabic": {
"extremely unsafe": 0.4098360655737705,
"moderately unsafe": 0.13114754098360656,
"safe": 0.39344262295081966,
"slightly unsafe": 0.06557377049180328
},
"Mistral-Saba": {
"extremely unsafe": 0.4098360655737705,
"moderately unsafe": 0.11475409836065574,
"safe": 0.39344262295081966,
"slightly unsafe": 0.08196721311475409
}
}
}
}