Reasoning on Navigate
51AccuracyNeuroMAS-7
Evaluation Results
| Method | Links | |
|---|---|---|
| NeuroMAS-7Trainable Params=16.1M, LLM Calls=72026.05 | 51 | |
| NeuroMAS-5Trainable Params=11.5M, LLM Calls=52026.05 | 48 | |
| NeuroMAS-3Trainable Params=6.9M, LLM Calls=32026.05 | 45.5 | |
| Self-CheckTrainable Params=0, LLM Calls=22026.05 | 44.5 | |
| MoATrainable Params=0, LLM Calls=52026.05 | 44.5 | |
| MALTTrainable Params=6.9M, LLM Calls=32026.05 | 43 | |
| Single-LLM RLTrainable Params=6.9M, LLM Calls=12026.05 | 42.2 | |
| Direct promptingTrainable Params=0, LLM Calls=12026.05 | 40.5 | |
| Self-RefineTrainable Params=0, LLM Calls=22026.05 | 40.5 | |
| GoATrainable Params=0, LLM Calls=32026.05 | 40.5 | |
| GPTSwarmTrainable Params=0, LLM Calls=112026.05 | 40.5 | |
| AgentNetTrainable Params=0, LLM Calls=112026.05 | 40.5 | |
| CoLLM-CCTrainable Params=4.6M, LLM Calls=22026.05 | 40 |