| name | ttc-router |
| description | TTC router skill for inference scaling decisions. Use when training, evaluating, or benchmarking the router that decides between low/medium/high/superthink compute modes. |
| compatibility | Requires router model and benchmark datasets |
TTC Router Skill
The TTC router decides which compute level to use for each query:
- low: Quick responses (~1s)
- medium: Standard reasoning (~5s)
- high: Deep thinking (~30s)
- superthink: Maximum compute (~2min)
Router Architecture
┌─────────────┐
│ Router │
│ Classifier │
└──────┬──────┘
│
┌────────────────┼────────────────┐
▼ ▼ ▼
┌──────────┐ ┌──────────┐ ┌──────────┐
│ LOW │ │ MEDIUM │ │ HIGH │
│ 1 sample │ │ 3 samples│ │ 10 samples│
└──────────┘ └──────────┘ └──────────┘
Router Training
python scripts/train_router.py \
--base-model Qwen/Qwen2.5-0.5B \
--trace-data path/to/traces.jsonl \
--output models/ttc-router \
--backend kaggle
Training Data Format
{"query": "What is 2+2?", "complexity": "low", "optimal_samples": 1}
{"query": "Solve this math problem...", "complexity": "high", "optimal_samples": 10}
Router Evaluation
python scripts/eval_router.py \
--router models/ttc-router \
--benchmark mmlu \
--modes low,medium,high
Evaluation Metrics
| Metric | Description |
|---|
| Accuracy | Correct routing decisions |
| Efficiency | Average compute per query |
| Coverage | Distribution across modes |
Router Benchmarks
| Benchmark | Low | Medium | High | Router Choice |
|---|
| MMLU | 45% | 58% | 62% | Mixed |
| GSM8K | 30% | 55% | 70% | Pred. High |
| Simple QA | 85% | 86% | 87% | Pred. Low |
Superthink Mode
For maximum reasoning:
python scripts/superthink.py \
--model Qwen/Qwen2.5-7B-Instruct \
--query "Complex problem..." \
--samples 50 \
--aggregation best_of_n
Example
User: Train a router that knows when to use more thinking
Agent will:
1. Collect inference traces with varying compute
2. Train classifier on complexity labels
3. Evaluate on held-out benchmark
4. Return router model
References
