Dashboard

Learning Engine

Self-evolving routing weights and insights from query outcomes.

Routing Weights

Cycle 847 / EMA 0.80

Inference

0.359 +0.041

Prediction

0.233 +0.018

Scraping

0.209 -0.012

Storage

0.144 +0.007

Annotation

0.049 -0.003

Training

0.006 +0.002

Overall Accuracy 74.2%

296 total

Cycle 847 positive

Multi-subnet queries involving SN1+SN13 produce 23% higher quality than SN1 alone for research tasks.

Cycle 845 caution

SN8 prediction accuracy drops below 60% during high-volatility network periods. Route to SN1 reasoning instead.

Cycle 842 positive

Adding SN33 metacognition step improves overall quality by 8% but adds 1.2s latency. Worth it for complex queries.

Cycle 839 caution

Real-time subnets (SN19) have 3x higher failure rate during peak hours. Implement fallback routing.

Cycle 836 insight

Optimal decomposition depth is 3-5 subtasks. Beyond 5, synthesis quality degrades due to context dilution.