File size: 25,571 Bytes
9a0283b | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 | #!/usr/bin/env python3
"""
Final Comprehensive Analysis Script
Analyzes the comprehensive test results: 256 models × 20 tests
Generates all figures and tables for the paper
"""
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns
from pathlib import Path
from datetime import datetime
import json
# Set style for publication-quality figures
plt.style.use('seaborn-v0_8-whitegrid')
sns.set_palette("husl")
def load_results():
"""Load the comprehensive test results."""
excel_file = 'comprehensive_20_tests_results_20251014_153008.xlsx'
json_file = 'comprehensive_20_tests_results_20251014_153008.json'
print("="*80)
print("LOADING COMPREHENSIVE TEST RESULTS")
print("="*80)
print(f"\nLoading from: {excel_file}")
# Load Excel file
xls = pd.ExcelFile(excel_file)
# Load all sheets
all_results = pd.read_excel(xls, 'All Results')
model_rankings = pd.read_excel(xls, 'Model Rankings', index_col=0)
test_difficulty = pd.read_excel(xls, 'Test Difficulty')
category_performance = pd.read_excel(xls, 'Category Performance', index_col=0)
print(f" Total results: {len(all_results)}")
print(f" Models tested: {all_results['model'].nunique()}")
print(f" Tests conducted: {all_results['test_id'].nunique()}")
# Load JSON for additional metadata
with open(json_file, 'r') as f:
json_data = json.load(f)
return all_results, model_rankings, test_difficulty, category_performance, json_data
def print_summary_statistics(all_results, json_data):
"""Print comprehensive summary statistics."""
print("\n" + "="*80)
print("SUMMARY STATISTICS")
print("="*80)
metadata = json_data['metadata']
summary = json_data['summary']
print(f"\nDataset Overview:")
print(f" Total Models: {metadata['total_models']}")
print(f" Total Tests: {metadata['total_tests']}")
print(f" Total Evaluations: {metadata['total_results']}")
print(f" Timestamp: {metadata['timestamp']}")
print(f"\nOverall Performance:")
print(f" Overall Pass Rate: {summary['overall_pass_rate']:.1f}%")
print(f" Best Model: {summary['best_model']} ({summary['best_model_score']:.1f}%)")
print(f" Hardest Test: Test {summary['hardest_test']} ({summary['hardest_test_pass_rate']:.1f}% pass rate)")
# API success rate
success_rate = (all_results['status'] == 'success').mean() * 100
print(f" API Success Rate: {success_rate:.1f}%")
# Response time statistics
avg_response_time = all_results[all_results['response_time'] > 0]['response_time'].mean()
print(f" Average Response Time: {avg_response_time:.2f}s")
def print_top_models(model_rankings):
"""Print top performing models."""
print("\n" + "="*80)
print("TOP 20 PERFORMING MODELS")
print("="*80)
print(f"\n{'Rank':<6} {'Pass Rate':<12} {'Model'}")
print("-" * 80)
for idx, (model, row) in enumerate(model_rankings.head(20).iterrows(), 1):
pass_rate = row['Pass Rate (%)']
print(f"{idx:<6} {pass_rate:>6.1f}% {model}")
def print_test_difficulty(test_difficulty):
"""Print test difficulty analysis."""
print("\n" + "="*80)
print("TEST DIFFICULTY ANALYSIS (HARDEST TO EASIEST)")
print("="*80)
print(f"\n{'ID':<4} {'Pass Rate':<12} {'Category':<25} {'Test Name'}")
print("-" * 95)
# Sort by pass rate (ascending = hardest first)
sorted_tests = test_difficulty.sort_values('Pass Rate (%)')
for _, row in sorted_tests.iterrows():
test_id = int(row['Test ID'])
pass_rate = row['Pass Rate (%)']
category = row['category'][:23] if pd.notna(row['category']) else ''
name = row['name'][:45]
print(f"{test_id:<4} {pass_rate:>6.1f}% {category:<25} {name}")
def print_category_analysis(category_performance):
"""Print category performance analysis."""
print("\n" + "="*80)
print("PERFORMANCE BY CATEGORY")
print("="*80)
# Sort by pass rate
sorted_cats = category_performance.sort_values('Pass Rate (%)')
print(f"\n{'Category':<30} {'Pass Rate'}")
print("-" * 45)
for category, row in sorted_cats.iterrows():
pass_rate = row['Pass Rate (%)']
print(f"{category:<30} {pass_rate:>6.1f}%")
def analyze_by_provider(all_results):
"""Analyze performance by model provider."""
print("\n" + "="*80)
print("PROVIDER ANALYSIS")
print("="*80)
# Extract provider from model name
all_results['provider'] = all_results['model'].apply(
lambda x: x.split('/')[0] if '/' in x else 'other'
)
# Calculate provider statistics
provider_stats = all_results.groupby('provider').agg({
'passed': 'mean',
'model': 'nunique',
'test_id': 'count'
}).round(3)
provider_stats.columns = ['pass_rate', 'num_models', 'total_tests']
provider_stats['pass_rate'] = provider_stats['pass_rate'] * 100
provider_stats = provider_stats.sort_values('pass_rate', ascending=False)
# Filter to providers with at least 3 models
provider_stats_filtered = provider_stats[provider_stats['num_models'] >= 3]
print(f"\n{'Provider':<20} {'Pass Rate':<12} {'Models':<10} {'Tests'}")
print("-" * 55)
for provider, row in provider_stats_filtered.head(15).iterrows():
print(f"{provider:<20} {row['pass_rate']:>6.1f}% {int(row['num_models']):<10} {int(row['total_tests'])}")
return provider_stats_filtered
def create_heatmap(all_results, output_file='fig1_heatmap.pdf'):
"""Create heatmap of model performance on all tests."""
print("\n" + "="*80)
print("CREATING FIGURE 1: PERFORMANCE HEATMAP (Top 50 Models)")
print("="*80)
# Create pivot table
pivot = all_results.pivot_table(
index='model',
columns='test_id',
values='passed',
aggfunc='first'
)
# Select top 50 models by overall performance
model_scores = pivot.mean(axis=1).sort_values(ascending=False)
top_50_models = model_scores.head(50).index
pivot_top50 = pivot.loc[top_50_models]
# Create figure
fig, ax = plt.subplots(figsize=(14, 12))
# Create heatmap
sns.heatmap(
pivot_top50,
cmap=['#d73027', '#91cf60'], # Red for fail, green for pass
cbar_kws={'label': 'Pass (1) / Fail (0)', 'ticks': [0, 1]},
xticklabels=True,
yticklabels=True,
vmin=0,
vmax=1,
linewidths=0.3,
linecolor='gray',
ax=ax
)
plt.title('Model Performance on 20 Diagnostic Tests (Top 50 Models)',
fontsize=16, fontweight='bold', pad=20)
plt.xlabel('Test ID', fontsize=12)
plt.ylabel('Model', fontsize=12)
plt.yticks(fontsize=7)
plt.xticks(fontsize=10)
plt.tight_layout()
plt.savefig(output_file, dpi=300, bbox_inches='tight')
print(f"✅ Saved: {output_file}")
plt.close()
def create_provider_chart(provider_stats, output_file='fig2_provider.pdf'):
"""Create bar chart of provider performance."""
print("\n" + "="*80)
print("CREATING FIGURE 2: PROVIDER COMPARISON")
print("="*80)
# Select top 12 providers
top_providers = provider_stats.head(12)
# Create figure
fig, ax = plt.subplots(figsize=(12, 7))
# Create bar plot
x_pos = np.arange(len(top_providers))
colors = sns.color_palette('husl', len(top_providers))
bars = ax.bar(x_pos, top_providers['pass_rate'], color=colors, edgecolor='black', linewidth=0.5)
# Customize plot
ax.set_xlabel('Model Provider', fontsize=13, fontweight='bold')
ax.set_ylabel('Average Pass Rate (%)', fontsize=13, fontweight='bold')
ax.set_title('Performance by Model Provider (≥3 models)',
fontsize=16, fontweight='bold', pad=20)
ax.set_xticks(x_pos)
ax.set_xticklabels(top_providers.index, rotation=45, ha='right', fontsize=11)
ax.set_ylim([0, max(top_providers['pass_rate']) * 1.15])
ax.grid(axis='y', alpha=0.3)
# Add value labels on bars
for bar in bars:
height = bar.get_height()
ax.text(bar.get_x() + bar.get_width()/2., height + 1,
f'{height:.1f}%', ha='center', va='bottom', fontsize=10, fontweight='bold')
# Add average line
avg_rate = top_providers['pass_rate'].mean()
ax.axhline(y=avg_rate, color='red', linestyle='--', alpha=0.7, linewidth=2,
label=f'Average: {avg_rate:.1f}%')
ax.legend(fontsize=11)
plt.tight_layout()
plt.savefig(output_file, dpi=300, bbox_inches='tight')
print(f"✅ Saved: {output_file}")
plt.close()
def create_difficulty_chart(test_difficulty, output_file='fig3_difficulty.pdf'):
"""Create horizontal bar chart of test difficulty."""
print("\n" + "="*80)
print("CREATING FIGURE 3: TEST DIFFICULTY RANKING")
print("="*80)
# Sort by difficulty (ascending pass rate = harder)
sorted_tests = test_difficulty.sort_values('Pass Rate (%)')
# Create shortened labels
labels = []
for _, row in sorted_tests.iterrows():
test_id = int(row['Test ID'])
name = row['name']
# Shorten long names
if len(name) > 35:
name = name[:32] + '...'
labels.append(f"T{test_id}: {name}")
pass_rates = sorted_tests['Pass Rate (%)'].values
# Create figure
fig, ax = plt.subplots(figsize=(12, 10))
# Color based on difficulty
colors = []
for rate in pass_rates:
if rate < 10:
colors.append('#8B0000') # Dark red - extremely hard
elif rate < 20:
colors.append('#d73027') # Red - very hard
elif rate < 40:
colors.append('#fdae61') # Orange - hard
elif rate < 60:
colors.append('#fee08b') # Yellow - medium
elif rate < 80:
colors.append('#a6d96a') # Light green - easy
else:
colors.append('#1a9850') # Dark green - very easy
# Create horizontal bar plot
y_pos = np.arange(len(labels))
bars = ax.barh(y_pos, pass_rates, color=colors, edgecolor='black', linewidth=0.5)
# Customize plot
ax.set_xlabel('Pass Rate (%)', fontsize=13, fontweight='bold')
ax.set_ylabel('Test', fontsize=13, fontweight='bold')
ax.set_title('Test Difficulty Ranking (Hardest to Easiest)',
fontsize=16, fontweight='bold', pad=20)
ax.set_yticks(y_pos)
ax.set_yticklabels(labels, fontsize=9)
ax.set_xlim([0, 105])
ax.grid(axis='x', alpha=0.3)
# Add value labels
for bar, rate in zip(bars, pass_rates):
width = bar.get_width()
ax.text(width + 1.5, bar.get_y() + bar.get_height()/2.,
f'{rate:.1f}%', ha='left', va='center', fontsize=9, fontweight='bold')
# Add vertical reference lines
ax.axvline(x=50, color='black', linestyle='--', alpha=0.3, linewidth=1)
ax.axvline(x=25, color='red', linestyle=':', alpha=0.3, linewidth=1)
ax.axvline(x=75, color='green', linestyle=':', alpha=0.3, linewidth=1)
# Add legend for difficulty colors
from matplotlib.patches import Patch
legend_elements = [
Patch(facecolor='#8B0000', label='Extremely Hard (<10%)'),
Patch(facecolor='#d73027', label='Very Hard (10-20%)'),
Patch(facecolor='#fdae61', label='Hard (20-40%)'),
Patch(facecolor='#fee08b', label='Medium (40-60%)'),
Patch(facecolor='#a6d96a', label='Easy (60-80%)'),
Patch(facecolor='#1a9850', label='Very Easy (>80%)')
]
ax.legend(handles=legend_elements, loc='lower right', fontsize=9)
plt.tight_layout()
plt.savefig(output_file, dpi=300, bbox_inches='tight')
print(f"✅ Saved: {output_file}")
plt.close()
def create_category_chart(category_performance, output_file='fig4_category.pdf'):
"""Create bar chart of category performance."""
print("\n" + "="*80)
print("CREATING FIGURE 4: CATEGORY PERFORMANCE")
print("="*80)
# Sort by pass rate
sorted_cats = category_performance.sort_values('Pass Rate (%)')
# Create figure
fig, ax = plt.subplots(figsize=(10, 6))
# Create bar plot
x_pos = np.arange(len(sorted_cats))
colors = plt.cm.RdYlGn(sorted_cats['Pass Rate (%)'] / 100)
bars = ax.bar(x_pos, sorted_cats['Pass Rate (%)'], color=colors, edgecolor='black', linewidth=0.8)
# Customize plot
ax.set_xlabel('Category', fontsize=13, fontweight='bold')
ax.set_ylabel('Pass Rate (%)', fontsize=13, fontweight='bold')
ax.set_title('Performance by Test Category',
fontsize=16, fontweight='bold', pad=20)
ax.set_xticks(x_pos)
ax.set_xticklabels(sorted_cats.index, rotation=45, ha='right', fontsize=11)
ax.set_ylim([0, 100])
ax.grid(axis='y', alpha=0.3)
# Add value labels
for bar in bars:
height = bar.get_height()
ax.text(bar.get_x() + bar.get_width()/2., height + 2,
f'{height:.1f}%', ha='center', va='bottom', fontsize=11, fontweight='bold')
plt.tight_layout()
plt.savefig(output_file, dpi=300, bbox_inches='tight')
print(f"✅ Saved: {output_file}")
plt.close()
def generate_latex_tables(model_rankings, test_difficulty, category_performance, provider_stats):
"""Generate LaTeX tables for the paper."""
print("\n" + "="*80)
print("GENERATING LATEX TABLES")
print("="*80)
output_file = 'paper_tables.tex'
with open(output_file, 'w') as f:
# Table 1: Top Models
f.write("% Table 1: Top 15 Performing Models\n")
f.write("\\begin{table}[htbp]\n")
f.write("\\centering\n")
f.write("\\caption{Top 15 Performing Models on 20 Diagnostic Tests}\n")
f.write("\\label{tab:top_models}\n")
f.write("\\begin{tabular}{rlr}\n")
f.write("\\toprule\n")
f.write("Rank & Model & Pass Rate \\\\\n")
f.write("\\midrule\n")
for idx, (model, row) in enumerate(model_rankings.head(15).iterrows(), 1):
model_escaped = model.replace('_', '\\_').replace('&', '\\&')
pass_rate = row['Pass Rate (%)']
f.write(f"{idx} & \\texttt{{{model_escaped}}} & {pass_rate:.1f}\\% \\\\\n")
f.write("\\bottomrule\n")
f.write("\\end{tabular}\n")
f.write("\\end{table}\n\n")
# Table 2: Test Difficulty
f.write("% Table 2: Test Difficulty (Hardest 10)\n")
f.write("\\begin{table}[htbp]\n")
f.write("\\centering\n")
f.write("\\caption{Test Difficulty Analysis (10 Hardest Tests)}\n")
f.write("\\label{tab:test_difficulty}\n")
f.write("\\begin{tabular}{clrl}\n")
f.write("\\toprule\n")
f.write("ID & Test Name & Pass Rate & Category \\\\\n")
f.write("\\midrule\n")
sorted_tests = test_difficulty.sort_values('Pass Rate (%)')
for _, row in sorted_tests.head(10).iterrows():
test_id = int(row['Test ID'])
name = row['name'][:40]
name_escaped = name.replace('_', '\\_').replace('&', '\\&')
pass_rate = row['Pass Rate (%)']
category = row['category'][:20] if pd.notna(row['category']) else ''
category_escaped = category.replace('_', '\\_').replace('&', '\\&')
f.write(f"{test_id} & {name_escaped} & {pass_rate:.1f}\\% & {category_escaped} \\\\\n")
f.write("\\bottomrule\n")
f.write("\\end{tabular}\n")
f.write("\\end{table}\n\n")
# Table 3: Category Performance
f.write("% Table 3: Category Performance\n")
f.write("\\begin{table}[htbp]\n")
f.write("\\centering\n")
f.write("\\caption{Performance by Test Category}\n")
f.write("\\label{tab:category_performance}\n")
f.write("\\begin{tabular}{lr}\n")
f.write("\\toprule\n")
f.write("Category & Pass Rate \\\\\n")
f.write("\\midrule\n")
sorted_cats = category_performance.sort_values('Pass Rate (%)')
for category, row in sorted_cats.iterrows():
category_escaped = category.replace('_', '\\_').replace('&', '\\&')
pass_rate = row['Pass Rate (%)']
f.write(f"{category_escaped} & {pass_rate:.1f}\\% \\\\\n")
f.write("\\bottomrule\n")
f.write("\\end{tabular}\n")
f.write("\\end{table}\n\n")
# Table 4: Provider Comparison
f.write("% Table 4: Provider Comparison (Top 10)\n")
f.write("\\begin{table}[htbp]\n")
f.write("\\centering\n")
f.write("\\caption{Performance by Model Provider}\n")
f.write("\\label{tab:provider_comparison}\n")
f.write("\\begin{tabular}{lrrr}\n")
f.write("\\toprule\n")
f.write("Provider & Models & Pass Rate & Tests \\\\\n")
f.write("\\midrule\n")
for provider, row in provider_stats.head(10).iterrows():
provider_escaped = provider.replace('_', '\\_').replace('&', '\\&')
num_models = int(row['num_models'])
pass_rate = row['pass_rate']
total_tests = int(row['total_tests'])
f.write(f"{provider_escaped} & {num_models} & {pass_rate:.1f}\\% & {total_tests} \\\\\n")
f.write("\\bottomrule\n")
f.write("\\end{tabular}\n")
f.write("\\end{table}\n")
print(f"✅ Saved: {output_file}")
def generate_summary_report(all_results, model_rankings, test_difficulty, category_performance, provider_stats, json_data):
"""Generate comprehensive summary report."""
print("\n" + "="*80)
print("GENERATING FINAL SUMMARY REPORT")
print("="*80)
output_file = 'FINAL_RESULTS_SUMMARY.md'
with open(output_file, 'w') as f:
f.write("# Comprehensive LLM Instruction Following Evaluation\n")
f.write(f"## Final Results Summary\n\n")
f.write(f"**Analysis Date:** {datetime.now().strftime('%Y-%m-%d %H:%M')}\n\n")
f.write(f"**Data Source:** {json_data['metadata']['timestamp']}\n\n")
f.write("---\n\n")
f.write("## Executive Summary\n\n")
metadata = json_data['metadata']
summary = json_data['summary']
f.write(f"This evaluation tested **{metadata['total_models']} large language models** ")
f.write(f"on **{metadata['total_tests']} diagnostic prompts**, ")
f.write(f"resulting in **{metadata['total_results']} individual test evaluations**.\n\n")
f.write(f"### Key Findings\n\n")
f.write(f"- **Overall Pass Rate:** {summary['overall_pass_rate']:.1f}%\n")
f.write(f"- **Best Performing Model:** {summary['best_model']} ({summary['best_model_score']:.1f}%)\n")
f.write(f"- **Hardest Test:** Test {summary['hardest_test']} ({summary['hardest_test_pass_rate']:.1f}% pass rate)\n")
# API success rate
success_rate = (all_results['status'] == 'success').mean() * 100
f.write(f"- **API Success Rate:** {success_rate:.1f}%\n\n")
f.write("---\n\n")
f.write("## Top 20 Performing Models\n\n")
f.write("| Rank | Model | Pass Rate |\n")
f.write("|------|-------|----------:|\n")
for idx, (model, row) in enumerate(model_rankings.head(20).iterrows(), 1):
pass_rate = row['Pass Rate (%)']
f.write(f"| {idx} | {model} | {pass_rate:.1f}% |\n")
f.write("\n---\n\n")
f.write("## Test Difficulty Analysis\n\n")
f.write("### Hardest Tests (Lowest Pass Rates)\n\n")
f.write("| ID | Test Name | Category | Pass Rate |\n")
f.write("|----|-----------|----------|----------:|\n")
sorted_tests = test_difficulty.sort_values('Pass Rate (%)')
for _, row in sorted_tests.head(10).iterrows():
test_id = int(row['Test ID'])
name = row['name']
category = row['category'] if pd.notna(row['category']) else 'N/A'
pass_rate = row['Pass Rate (%)']
f.write(f"| {test_id} | {name} | {category} | {pass_rate:.1f}% |\n")
f.write("\n### Easiest Tests (Highest Pass Rates)\n\n")
f.write("| ID | Test Name | Category | Pass Rate |\n")
f.write("|----|-----------|----------|----------:|\n")
for _, row in sorted_tests.tail(10).iterrows():
test_id = int(row['Test ID'])
name = row['name']
category = row['category'] if pd.notna(row['category']) else 'N/A'
pass_rate = row['Pass Rate (%)']
f.write(f"| {test_id} | {name} | {category} | {pass_rate:.1f}% |\n")
f.write("\n---\n\n")
f.write("## Performance by Category\n\n")
f.write("| Category | Pass Rate |\n")
f.write("|----------|----------:|\n")
sorted_cats = category_performance.sort_values('Pass Rate (%)')
for category, row in sorted_cats.iterrows():
pass_rate = row['Pass Rate (%)']
f.write(f"| {category} | {pass_rate:.1f}% |\n")
f.write("\n**Key Insight:** String manipulation tests are by far the hardest category, ")
f.write("while constraint compliance tests are the easiest.\n\n")
f.write("---\n\n")
f.write("## Performance by Provider\n\n")
f.write("Top providers (with ≥3 models):\n\n")
f.write("| Provider | Models | Pass Rate | Total Tests |\n")
f.write("|----------|--------|-----------|------------:|\n")
for provider, row in provider_stats.head(15).iterrows():
num_models = int(row['num_models'])
pass_rate = row['pass_rate']
total_tests = int(row['total_tests'])
f.write(f"| {provider} | {num_models} | {pass_rate:.1f}% | {total_tests} |\n")
f.write("\n---\n\n")
f.write("## Statistical Insights\n\n")
# Calculate additional statistics
model_pass_rates = all_results.groupby('model')['passed'].mean() * 100
test_pass_rates = all_results.groupby('test_id')['passed'].mean() * 100
f.write(f"### Model Performance Distribution\n\n")
f.write(f"- **Mean Pass Rate:** {model_pass_rates.mean():.1f}%\n")
f.write(f"- **Median Pass Rate:** {model_pass_rates.median():.1f}%\n")
f.write(f"- **Standard Deviation:** {model_pass_rates.std():.1f}%\n")
f.write(f"- **Models with 100% Pass Rate:** {(model_pass_rates == 100).sum()}\n")
f.write(f"- **Models with 0% Pass Rate:** {(model_pass_rates == 0).sum()}\n\n")
f.write(f"### Test Difficulty Distribution\n\n")
f.write(f"- **Mean Test Pass Rate:** {test_pass_rates.mean():.1f}%\n")
f.write(f"- **Median Test Pass Rate:** {test_pass_rates.median():.1f}%\n")
f.write(f"- **Standard Deviation:** {test_pass_rates.std():.1f}%\n")
f.write(f"- **Tests with >80% Pass Rate:** {(test_pass_rates > 80).sum()}\n")
f.write(f"- **Tests with <20% Pass Rate:** {(test_pass_rates < 20).sum()}\n\n")
f.write("---\n\n")
f.write("## Files Generated\n\n")
f.write("- `fig1_heatmap.pdf` - Performance heatmap (top 50 models)\n")
f.write("- `fig2_provider.pdf` - Provider comparison chart\n")
f.write("- `fig3_difficulty.pdf` - Test difficulty ranking\n")
f.write("- `fig4_category.pdf` - Category performance chart\n")
f.write("- `paper_tables.tex` - LaTeX tables for paper\n")
f.write("- `FINAL_RESULTS_SUMMARY.md` - This summary document\n\n")
f.write("---\n\n")
f.write("## Conclusions\n\n")
f.write("1. **Model Performance Varies Widely:** Pass rates range from 0% to 100%, ")
f.write("indicating significant differences in instruction-following capabilities.\n\n")
f.write("2. **String Manipulation is Hardest:** Tests requiring precise string manipulation ")
f.write("have the lowest pass rates, suggesting this is a key challenge for current LLMs.\n\n")
f.write("3. **Provider Differences:** Significant variation exists between model providers, ")
f.write("with top providers achieving much higher pass rates.\n\n")
f.write("4. **Few Perfect Models:** Only a small number of models achieve 100% pass rate, ")
f.write("indicating that even top models struggle with some tests.\n\n")
f.write("5. **API Reliability:** High API success rate indicates robust testing methodology.\n\n")
print(f"✅ Saved: {output_file}")
def main():
"""Main analysis function."""
print("\n" + "="*80)
print("COMPREHENSIVE FINAL ANALYSIS")
print(f"Analysis Date: {datetime.now().strftime('%Y-%m-%d %H:%M')}")
print("="*80 + "\n")
# Load results
all_results, model_rankings, test_difficulty, category_performance, json_data = load_results()
# Print summary statistics
print_summary_statistics(all_results, json_data)
print_top_models(model_rankings)
print_test_difficulty(test_difficulty)
print_category_analysis(category_performance)
# Analyze by provider
provider_stats = analyze_by_provider(all_results)
# Create all visualizations
create_heatmap(all_results)
create_provider_chart(provider_stats)
create_difficulty_chart(test_difficulty)
create_category_chart(category_performance)
# Generate LaTeX tables
generate_latex_tables(model_rankings, test_difficulty, category_performance, provider_stats)
# Generate summary report
generate_summary_report(all_results, model_rankings, test_difficulty, category_performance, provider_stats, json_data)
print("\n" + "="*80)
print("✅ COMPREHENSIVE ANALYSIS COMPLETE")
print("="*80)
print("\nAll figures, tables, and reports have been generated successfully!")
if __name__ == "__main__":
main()
|