Navigation Instruction Generation on R2R-Goal real-world (GO Stanford, ReCon, HuRoN)
0.27BLEU-4Anole-7B + Interleaved
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| Anole-7B + InterleavedEvaluation Protocol=Zero-shot, Base Model=Anole-7B, Reasoning Strategy=Interleaved2025.08 | 0.27 | 0.15 | 0.19 | 0.2 | |
| Anole-7B + One-passEvaluation Protocol=Zero-shot, Base Model=Anole-7B, Reasoning Strategy=One-pass2025.08 | 0.24 | 0.14 | 0.17 | 0.19 | |
| C-InstructorEvaluation Protocol=Zero-shot2025.08 | 0.15 | 0.08 | 0.12 | 0.15 | |
| GPT-4o + CoTEvaluation Protocol=Zero-shot, Reasoning Strategy=CoT2025.08 | 0.09 | 0.13 | 0.16 | 0.18 | |
| Claude 4 OpusEvaluation Protocol=Zero-shot2025.08 | 0.09 | 0.13 | 0.16 | 0.16 | |
| GPT-4oEvaluation Protocol=Zero-shot2025.08 | 0.08 | 0.11 | 0.18 | 0.17 | |
| Gemini 3.0Evaluation Protocol=Zero-shot2025.08 | 0.08 | 0.11 | 0.15 | 0.14 | |
| Anole-7B + CoTEvaluation Protocol=Zero-shot, Base Model=Anole-7B, Reasoning Strategy=CoT2025.08 | 0.08 | 0.1 | 0.13 | 0.17 | |
| Anole-7B + DirectEvaluation Protocol=Zero-shot, Base Model=Anole-7B, Reasoning Strategy=Direct2025.08 | 0.06 | 0.09 | 0.1 | 0.12 | |
| LANAEvaluation Protocol=Zero-shot2025.08 | 0.05 | 0.03 | 0.09 | 0.09 |