Reasoning explanation generation on ConversationGoT-120h (test)
4.46AlignmentGPT-5 (thinking)
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| GPT-5 (thinking)Latency (s)=16.98 ± 5.292026.02 | 4.46 | 4.33 | 4.32 | 4.65 | |
| Graph-of-ThoughtsLatency (s)=0.74 ± 0.122026.02 | 4.4 | 4.27 | 4.21 | 4.38 | |
| GPT-4oLatency (s)=2.98 ± 1.042026.02 | 3.4 | 3.27 | 3.21 | 3.38 | |
| Random selectorLatency (s)=0.73 ± 0.112026.02 | 3.28 | 3.13 | 3.21 | 3.88 |