Tracking LLM 's预测准确率，AnalyzeModel漂移趋势。Records AI 's每一个预测，验证其可靠性， Quantitatively evaluating different models' performance differences on prediction tasks.
Total Predictions
3
Accuracy
50%
Verified
2
Correct ✅
1
Error ❌
1
📊 Model Accuracy Comparison
GPT-4o
0% (0/0)
1 Pending
Claude 3.5
100% (1/1)
0 Pending
Gemini 2.0
0% (0/1)
0 Pending
⏳ PendingGPT-4oConfidence: 75%
Tesla stock will reach $350 in Q1 2026
💡 Based on FSD progress and earnings expectations
Category
Stocks
GoalDate
2026-03-31
Recorded on 2026-01-15
✅ CorrectClaude 3.5Confidence: 68%
The Fed will cut rates by 25 basis points in March 2026
💡 Inflation Data stabilizing
Category
Economy
GoalDate
2026-03-20
Recorded on 2026-02-01
❌ ErrorGemini 2.0Confidence: 55%
Apple will release AR glasses in Spring 2026
💡 Cook hints at delay to Fall 2026
Category
Product
GoalDate
2026-03-15
Recorded on 2026-01-20
🔬 Why Track LLM Predictions?
Large Language Models' prediction capabilities exhibit significant "model drift" — the accuracy of the same model may fluctuate across different time periods. Through long-term tracking, we can:
- Quantify the reliability of different models on prediction tasks
- Discover whether models have "overconfidence" issues
- Track changes in prediction capabilities after model updates
- Provide data support for decision-making, rather than blindly trusting AI