Action Learning on ALFRED (Q1)
92.5AccuracyCAG + Direct
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| CAG + DirectAvg Latency=1027.44 ms, Avg Output Tokens=02026.05 | 92.5 | 92.42 | 92.6 | 92.51 | |
| Affordance Routing (REIS)Avg Latency=406.21 ms, Avg Output Tokens=02026.05 | 92.5 | 92.08 | 93 | 92.54 | |
| RAG + DirectAvg Latency=1017.83 ms, Avg Output Tokens=02026.05 | 92.4 | 91.73 | 93.2 | 92.46 | |
| Base + DirectAvg Latency=327.56 ms, Avg Output Tokens=02026.05 | 39.7 | 39.55 | 39 | 39.27 | |
| Base + ReasoningAvg Latency=43.59 sec, Avg Output Tokens=219.82026.05 | 32.8 | 100 | 32.8 | 49.3 |