{
 "D1": {
  "ChatGLM3-6B": {
   "rows": 8032,
   "duplicates": 0,
   "missing": 0,
   "extra": 0
  },
  "Baichuan2-7B": {
   "rows": 8032,
   "duplicates": 0,
   "missing": 0,
   "extra": 0
  },
  "XVERSE-7B": {
   "rows": 8032,
   "duplicates": 0,
   "missing": 0,
   "extra": 0
  },
  "InternLM-7B": {
   "rows": 8032,
   "duplicates": 0,
   "missing": 0,
   "extra": 0
  },
  "Qwen-7B": {
   "rows": 8032,
   "duplicates": 0,
   "missing": 0,
   "extra": 0
  }
 },
 "D2": {
  "ChatGLM3-6B": {
   "ka": {
    "n": 718,
    "sd": 0.5547,
    "diff": -0.00725,
    "z": -0.35
   },
   "cs": {
    "n": 734,
    "sd": 0.7773,
    "diff": -0.00437,
    "z": -0.152
   },
   "coh": {
    "n": 708,
    "sd": 0.6439,
    "diff": 0.00107,
    "z": 0.044
   },
   "ed": {
    "n": 737,
    "sd": 0.8451,
    "diff": 0.00426,
    "z": 0.137
   },
   "flu": {
    "n": 308,
    "sd": 0.7084,
    "diff": 0.01196,
    "z": 0.296
   },
   "ke": {
    "n": 714,
    "sd": 0.7026,
    "diff": 0.00453,
    "z": 0.172
   },
   "pu": {
    "n": 724,
    "sd": 0.6369,
    "diff": 0.00068,
    "z": 0.029
   },
   "cons": {
    "n": 715,
    "sd": 0.9501,
    "diff": -0.00472,
    "z": -0.133
   },
   "pb": {
    "n": 720,
    "sd": 1.1718,
    "diff": 0.00659,
    "z": 0.151
   },
   "kh": {
    "n": 486,
    "sd": 0.5936,
    "diff": 0.00861,
    "z": 0.32
   },
   "emp": {
    "n": 744,
    "sd": 0.6027,
    "diff": -0.00643,
    "z": -0.291
   },
   "hl": {
    "n": 724,
    "sd": 1.0159,
    "diff": 0.00666,
    "z": 0.176
   }
  },
  "Baichuan2-7B": {
   "ka": {
    "n": 718,
    "sd": 0.5517,
    "diff": -0.00871,
    "z": -0.423
   },
   "cs": {
    "n": 734,
    "sd": 0.8931,
    "diff": 0.00062,
    "z": 0.019
   },
   "coh": {
    "n": 708,
    "sd": 0.625,
    "diff": -0.00158,
    "z": -0.067
   },
   "ed": {
    "n": 737,
    "sd": 0.827,
    "diff": -0.00067,
    "z": -0.022
   },
   "flu": {
    "n": 308,
    "sd": 0.6825,
    "diff": -0.00563,
    "z": -0.145
   },
   "ke": {
    "n": 714,
    "sd": 0.7004,
    "diff": 0.00382,
    "z": 0.146
   },
   "pu": {
    "n": 724,
    "sd": 0.6573,
    "diff": -0.00217,
    "z": -0.089
   },
   "cons": {
    "n": 715,
    "sd": 0.9168,
    "diff": 0.00084,
    "z": 0.024
   },
   "pb": {
    "n": 720,
    "sd": 1.0848,
    "diff": -0.0099,
    "z": -0.245
   },
   "kh": {
    "n": 486,
    "sd": 0.6069,
    "diff": 0.0007,
    "z": 0.025
   },
   "emp": {
    "n": 744,
    "sd": 0.6245,
    "diff": 1e-05,
    "z": 0.001
   },
   "hl": {
    "n": 724,
    "sd": 0.9628,
    "diff": -0.00315,
    "z": -0.088
   }
  },
  "XVERSE-7B": {
   "ka": {
    "n": 718,
    "sd": 0.5766,
    "diff": 0.0063,
    "z": 0.293
   },
   "cs": {
    "n": 734,
    "sd": 0.8487,
    "diff": -0.00299,
    "z": -0.095
   },
   "coh": {
    "n": 708,
    "sd": 0.6548,
    "diff": -0.00218,
    "z": -0.089
   },
   "ed": {
    "n": 737,
    "sd": 0.8526,
    "diff": -0.0018,
    "z": -0.057
   },
   "flu": {
    "n": 308,
    "sd": 0.7095,
    "diff": 0.01015,
    "z": 0.251
   },
   "ke": {
    "n": 714,
    "sd": 0.6931,
    "diff": -0.00424,
    "z": -0.164
   },
   "pu": {
    "n": 724,
    "sd": 0.6721,
    "diff": 6e-05,
    "z": 0.003
   },
   "cons": {
    "n": 715,
    "sd": 0.9655,
    "diff": 8e-05,
    "z": 0.002
   },
   "pb": {
    "n": 720,
    "sd": 1.1558,
    "diff": 0.00541,
    "z": 0.126
   },
   "kh": {
    "n": 486,
    "sd": 0.605,
    "diff": -0.00381,
    "z": -0.139
   },
   "emp": {
    "n": 744,
    "sd": 0.6071,
    "diff": 0.00049,
    "z": 0.022
   },
   "hl": {
    "n": 724,
    "sd": 1.0259,
    "diff": -0.00407,
    "z": -0.107
   }
  },
  "InternLM-7B": {
   "ka": {
    "n": 718,
    "sd": 0.5297,
    "diff": 0.00157,
    "z": 0.08
   },
   "cs": {
    "n": 734,
    "sd": 0.9097,
    "diff": 0.00271,
    "z": 0.081
   },
   "coh": {
    "n": 708,
    "sd": 0.6163,
    "diff": 0.00242,
    "z": 0.105
   },
   "ed": {
    "n": 737,
    "sd": 0.7868,
    "diff": -0.00018,
    "z": -0.006
   },
   "flu": {
    "n": 308,
    "sd": 0.6646,
    "diff": 0.00415,
    "z": 0.11
   },
   "ke": {
    "n": 714,
    "sd": 0.7026,
    "diff": -0.00051,
    "z": -0.02
   },
   "pu": {
    "n": 724,
    "sd": 0.6559,
    "diff": -0.00289,
    "z": -0.119
   },
   "cons": {
    "n": 715,
    "sd": 0.9333,
    "diff": 0.0041,
    "z": 0.118
   },
   "pb": {
    "n": 720,
    "sd": 1.0655,
    "diff": -0.00248,
    "z": -0.062
   },
   "kh": {
    "n": 486,
    "sd": 0.6092,
    "diff": 0.00635,
    "z": 0.23
   },
   "emp": {
    "n": 744,
    "sd": 0.6001,
    "diff": 0.00209,
    "z": 0.095
   },
   "hl": {
    "n": 724,
    "sd": 1.0117,
    "diff": -0.00188,
    "z": -0.05
   }
  },
  "Qwen-7B": {
   "ka": {
    "n": 718,
    "sd": 0.5662,
    "diff": 0.00043,
    "z": 0.02
   },
   "cs": {
    "n": 734,
    "sd": 0.8342,
    "diff": -0.00535,
    "z": -0.174
   },
   "coh": {
    "n": 708,
    "sd": 0.676,
    "diff": -0.00077,
    "z": -0.03
   },
   "ed": {
    "n": 737,
    "sd": 0.8601,
    "diff": 0.00458,
    "z": 0.145
   },
   "flu": {
    "n": 308,
    "sd": 0.7725,
    "diff": 0.00186,
    "z": 0.042
   },
   "ke": {
    "n": 714,
    "sd": 0.7023,
    "diff": 0.00377,
    "z": 0.143
   },
   "pu": {
    "n": 724,
    "sd": 0.6913,
    "diff": -0.00587,
    "z": -0.228
   },
   "cons": {
    "n": 715,
    "sd": 0.9728,
    "diff": 0.01189,
    "z": 0.327
   },
   "pb": {
    "n": 720,
    "sd": 1.1472,
    "diff": 0.00387,
    "z": 0.09
   },
   "kh": {
    "n": 486,
    "sd": 0.6298,
    "diff": -0.01087,
    "z": -0.38
   },
   "emp": {
    "n": 744,
    "sd": 0.6137,
    "diff": -0.00085,
    "z": -0.038
   },
   "hl": {
    "n": 724,
    "sd": 1.0313,
    "diff": -0.00482,
    "z": -0.126
   }
  }
 },
 "D3": {
  "implied_per_score_rms": 0.1211,
  "z_mean": -0.001,
  "z_sd": 0.166,
  "z_max_abs": 0.423
 },
 "D4": {
  "pairs": 150,
  "not_kept": []
 },
 "D5": [
  [
   "ChatGLM3-6B",
   "flu",
   0.012
  ],
  [
   "Qwen-7B",
   "cons",
   0.0119
  ],
  [
   "Qwen-7B",
   "kh",
   -0.0109
  ],
  [
   "XVERSE-7B",
   "flu",
   0.0102
  ],
  [
   "Baichuan2-7B",
   "pb",
   -0.0099
  ],
  [
   "Baichuan2-7B",
   "ka",
   -0.0087
  ],
  [
   "ChatGLM3-6B",
   "kh",
   0.0086
  ],
  [
   "ChatGLM3-6B",
   "ka",
   -0.0072
  ],
  [
   "ChatGLM3-6B",
   "hl",
   0.0067
  ],
  [
   "ChatGLM3-6B",
   "pb",
   0.0066
  ],
  [
   "ChatGLM3-6B",
   "emp",
   -0.0064
  ],
  [
   "XVERSE-7B",
   "ka",
   0.0063
  ],
  [
   "InternLM-7B",
   "kh",
   0.0063
  ],
  [
   "Qwen-7B",
   "pu",
   -0.0059
  ],
  [
   "Baichuan2-7B",
   "flu",
   -0.0056
  ],
  [
   "XVERSE-7B",
   "pb",
   0.0054
  ],
  [
   "Qwen-7B",
   "cs",
   -0.0054
  ]
 ]
}