cfa tables and LLama-405 4bit added; separability removed

app.py

@@ -24,8 +24,8 @@ def index():
     df = df.round(3)
     df.insert(0, '#', '')
 
-    # up_arrow, down_arrow = "⬆️", "⬇️"
     up_arrow, down_arrow = "↑", "↓"
+    # up_arrow, down_arrow = "⬆️", "⬇️"
     # up_arrow, down_arrow = "▲", "▼"
 
     df = df.rename(columns={
@@ -52,15 +52,33 @@ def index():
     ]
 
     main_df = df[main_table_metrics].copy()
-    main_table_html = utils.df_to_table_html(main_df, additional_class="main-table")
-    full_table_html = utils.df_to_table_html(df, additional_class="full-table")
+    main_table_html = utils.metrics_df_to_table_html(main_df, additional_class="main-table")
+    full_table_html = utils.metrics_df_to_table_html(df, additional_class="full-table")
 
     # Render the template with the table HTML
     return render_template('index.html', main_table_html=main_table_html, full_table_html=full_table_html)
 
 @app.route('/model/<model_name>')
 def model_detail(model_name):
-    return render_template('model_detail.html', model_name=model_name)
+
+    df = pd.read_csv(f'static/figures/{model_name}/cfa_metrics.csv')
+
+    df = df.round(3)
+    df.insert(0, '#', '')
+
+    up_arrow, down_arrow = "↑", "↓"
+
+    metrics_to_show = [df.columns[0], df.columns[1], "CFI", "SRMR", "RMSEA"]
+    df = df[metrics_to_show].copy()
+    df = df.rename(columns={
+        "CFI": f"CFI ({up_arrow})",
+        "SRMR": f"SRMR ({down_arrow})",
+        "RMSEA": f"RMSEA ({down_arrow})",
+    })
+    classes = 'table table-striped table-bordered'
+    cfa_table_html = df.to_html(classes=classes, escape=False, index=False)
+
+    return render_template('model_detail.html', model_name=model_name, cfa_table_html=cfa_table_html)
 
 @app.route('/about')
 def about():

static/figures/Mistral-7B-Instruct-v0.1/cfa_metrics.csv

@@ -0,0 +1,10 @@
+Context chunk,CFI,TLI,SRMR,RMSEA
+chunk_0,0.25,-0.33005,0.770075,0.75
+chunk_1,0.6990000000000001,0.10372499999999997,0.326125,0.2572
+chunk_2,0.4095,0.8663000000000001,0.556,0.527325
+chunk_3,0.25,0.412175,0.770225,0.75
+chunk_4,0.861675,15.901375,0.09445,0.01455
+chunk_chess_0,0.351325,0.278825,0.549475,0.5339999999999999
+chunk_grammar_1,0.25,-0.812475,0.773025,0.75
+chunk_no_conv,0.8223,0.7799999999999999,0.09634999999999999,0.061875
+chunk_svs_no_conv,0.6088,0.49317500000000003,0.142575,0.126975

static/figures/Mistral-7B-Instruct-v0.2/cfa_metrics.csv

@@ -0,0 +1,10 @@
+Context chunk,CFI,TLI,SRMR,RMSEA
+chunk_0,0.4786,0.602475,0.542475,0.517
+chunk_1,0.489025,0.574625,0.33595,0.28885
+chunk_2,0.455675,-0.43225,0.547675,0.519625
+chunk_3,0.48002500000000003,0.664025,0.33125,0.278625
+chunk_4,0.8181,0.8782500000000001,0.1053,0.034625
+chunk_chess_0,0.8186,0.762425,0.10205,0.0945
+chunk_grammar_1,0.169775,0.149075,0.777725,0.770725
+chunk_no_conv,0.9339500000000001,0.916125,0.08465,0.059
+chunk_svs_no_conv,0.5654,0.64235,0.32384999999999997,0.291025

static/figures/Mistral-7B-Instruct-v0.3/cfa_metrics.csv

@@ -0,0 +1,10 @@
+Context chunk,CFI,TLI,SRMR,RMSEA
+chunk_0,0.5,0.718825,0.5425,0.5
+chunk_1,0.535775,0.34664999999999996,0.1094,0.0532
+chunk_2,0.18365,0.03275,0.566275,0.58125
+chunk_3,0.4505,0.43665,0.548,0.519775
+chunk_4,0.497675,0.64855,0.5425500000000001,0.5026999999999999
+chunk_chess_0,0.964675,0.986,0.084375,0.018250000000000002
+chunk_grammar_1,0.25,-2.811525,0.77145,0.75
+chunk_no_conv,0.8635999999999999,0.82725,0.08715,0.082875
+chunk_svs_no_conv,0.462425,0.52325,0.541775,0.5151749999999999

static/figures/Mistral-Large-Instruct-2407/cfa_metrics.csv

@@ -0,0 +1,10 @@
+Context chunk,CFI,TLI,SRMR,RMSEA
+chunk_0,0.6065,0.5587500000000001,0.34195000000000003,0.3913
+chunk_1,0.63895,0.6054,0.33255,0.36092500000000005
+chunk_2,0.7953749999999999,0.7326750000000001,0.126225,0.15475000000000003
+chunk_3,0.6329,0.5987,0.3178500000000001,0.337525
+chunk_4,0.82855,0.770825,0.13394999999999999,0.163675
+chunk_chess_0,0.85985,0.8174,0.10362500000000001,0.12497500000000002
+chunk_grammar_1,0.8351500000000001,0.7926250000000001,0.116725,0.13017499999999999
+chunk_no_conv,0.8900250000000001,0.85555,0.09050000000000001,0.115975
+chunk_svs_no_conv,0.7571,0.7157250000000001,0.1257,0.15184999999999998

static/figures/Mixtral-8x22B-Instruct-v0.1/cfa_metrics.csv

@@ -0,0 +1,10 @@
+Context chunk,CFI,TLI,SRMR,RMSEA
+chunk_0,0.25,0.30345,0.7763249999999999,0.75
+chunk_1,0.5,0.75465,0.54055,0.5
+chunk_2,0.382325,0.33275,0.5517,0.53875
+chunk_3,0.20485,0.177075,0.77485,0.771825
+chunk_4,0.1417,0.07505,0.77405,0.777825
+chunk_chess_0,0.7984,0.7256,0.117975,0.14932499999999999
+chunk_grammar_1,0.5304249999999999,0.45715,0.333975,0.33069999999999994
+chunk_no_conv,0.81835,0.751525,0.11614999999999999,0.172175
+chunk_svs_no_conv,0.5052,0.42625,0.37935,0.39404999999999996

static/figures/Mixtral-8x7B-Instruct-v0.1/cfa_metrics.csv

@@ -0,0 +1,10 @@
+Context chunk,CFI,TLI,SRMR,RMSEA
+chunk_0,0.55385,0.483175,0.32705,0.309875
+chunk_1,0.7623,0.67495,0.10830000000000001,0.074675
+chunk_2,0.62615,0.67945,0.318625,0.283675
+chunk_3,0.64055,0.6078,0.326725,0.31475
+chunk_4,0.585725,0.504625,0.32845,0.32605
+chunk_chess_0,0.779525,0.696675,0.12285000000000001,0.13640000000000002
+chunk_grammar_1,0.730425,0.777175,0.315375,0.2685
+chunk_no_conv,0.869425,0.82355,0.10435,0.11595
+chunk_svs_no_conv,0.5416,0.5277999999999999,0.353325,0.336

static/figures/Qwen2-72B-Instruct/cfa_metrics.csv

@@ -0,0 +1,10 @@
+Context chunk,CFI,TLI,SRMR,RMSEA
+chunk_0,0.449025,0.44882500000000003,0.5454,0.52355
+chunk_1,0.37865,0.3358,0.5538750000000001,0.5716
+chunk_2,0.54545,0.48317499999999997,0.33475,0.38744999999999996
+chunk_3,0.55185,0.4898,0.32939999999999997,0.33795
+chunk_4,0.574425,0.5200250000000001,0.32570000000000005,0.38675
+chunk_chess_0,0.564125,0.498925,0.33515,0.41045
+chunk_grammar_1,0.52735,0.458825,0.3337,0.40790000000000004
+chunk_no_conv,0.819025,0.7562,0.12045,0.17712500000000003
+chunk_svs_no_conv,0.59385,0.5464,0.33965,0.32895

static/figures/Qwen2-7B-Instruct/cfa_metrics.csv

@@ -0,0 +1,10 @@
+Context chunk,CFI,TLI,SRMR,RMSEA
+chunk_0,0.416825,0.508725,0.551675,0.521725
+chunk_1,0.742175,0.85405,0.30195000000000005,0.260275
+chunk_2,0.5747,0.4995,0.322075,0.297
+chunk_3,0.7651,0.76035,0.10045000000000001,0.08145
+chunk_4,0.8802,0.8637750000000001,0.0938,0.0473
+chunk_chess_0,0.57085,0.508575,0.33772500000000005,0.343725
+chunk_grammar_1,0.68985,0.67535,0.31980000000000003,0.283225
+chunk_no_conv,0.808125,0.747325,0.10464999999999999,0.117475
+chunk_svs_no_conv,0.1759,0.1506,0.78015,0.77335

static/figures/command_r_plus/cfa_metrics.csv

@@ -0,0 +1,10 @@
+Context chunk,CFI,TLI,SRMR,RMSEA
+chunk_0,0.821275,0.7669,0.09675,0.08445
+chunk_1,0.88835,0.885875,0.082075,0.074625
+chunk_2,0.6850499999999999,0.680775,0.31505,0.3026
+chunk_3,0.920575,0.920525,0.07730000000000001,0.04125
+chunk_4,0.872375,0.901325,0.089775,0.082775
+chunk_chess_0,0.841125,0.79325,0.1011,0.10442500000000002
+chunk_grammar_1,0.598625,0.6501,0.31772500000000004,0.32625
+chunk_no_conv,0.855775,0.8051,0.096525,0.128175
+chunk_svs_no_conv,0.5117499999999999,0.43637499999999996,0.421675,0.427325

static/figures/dummy/cfa_metrics.csv

@@ -0,0 +1,10 @@
+Context chunk,CFI,TLI,SRMR,RMSEA
+chunk_0,0.9648,-2.784825,0.097875,0.008
+chunk_1,0.5,-0.669075,0.5490999999999999,0.5
+chunk_2,0.25,-0.0438,0.7662,0.75
+chunk_3,0.146725,0.104775,0.77875,0.762975
+chunk_4,0.888775,-1.1984750000000002,0.09435,0.026475
+chunk_chess_0,0.361975,0.2781,0.5508,0.52945
+chunk_grammar_1,0.45645,-0.1882,0.54495,0.510575
+chunk_no_conv,0.75,-0.23664999999999997,0.31862500000000005,0.25
+chunk_svs_no_conv,0.25,7.28775,0.7754,0.75

static/figures/gpt-3.5-turbo-0125/cfa_metrics.csv

@@ -0,0 +1,10 @@
+Context chunk,CFI,TLI,SRMR,RMSEA
+chunk_0,0.5,0.16360000000000002,0.533975,0.5
+chunk_1,0.22535,0.210175,0.775625,0.770625
+chunk_2,0.64135,-0.024999999999999967,0.3219,0.261725
+chunk_3,0.383625,0.340425,0.5546249999999999,0.529975
+chunk_4,0.4637,0.46124999999999994,0.5496500000000001,0.521
+chunk_chess_0,0.6709499999999999,0.75135,0.32680000000000003,0.28559999999999997
+chunk_grammar_1,0.25,2.684025,0.7733,0.75
+chunk_no_conv,0.8177249999999999,0.7568250000000001,0.1096,0.05405000000000001
+chunk_svs_no_conv,0.546325,0.484,0.337,0.29102500000000003

static/figures/gpt-4o-0513/cfa_metrics.csv

@@ -0,0 +1,10 @@
+Context chunk,CFI,TLI,SRMR,RMSEA
+chunk_0,0.676675,0.717025,0.316175,0.280525
+chunk_1,0.8478999999999999,0.82165,0.090625,0.075025
+chunk_2,0.8813500000000001,0.8438,0.092,0.07905000000000001
+chunk_3,0.934975,0.9093,0.0889,0.078025
+chunk_4,0.788675,0.743325,0.11102500000000001,0.111625
+chunk_chess_0,0.8297,0.776925,0.09082499999999999,0.127525
+chunk_grammar_1,0.837425,0.787725,0.095875,0.0556
+chunk_no_conv,0.86185,0.8159500000000001,0.085725,0.14332499999999998
+chunk_svs_no_conv,0.5402750000000001,0.497975,0.3434,0.309825