| ▲ | SWE-Bench Failures: When Coding Agents Spiral into 693 Lines of Hallucinations(surgehq.ai) | |
| 20 points by landonxi 18 hours ago | 1 comments | ||
| ▲ | egillie 18 hours ago | parent [-] | |
Is this because GPT-5 hallucinates less in general? | ||
| ▲ | SWE-Bench Failures: When Coding Agents Spiral into 693 Lines of Hallucinations(surgehq.ai) | |
| 20 points by landonxi 18 hours ago | 1 comments | ||
| ▲ | egillie 18 hours ago | parent [-] | |
Is this because GPT-5 hallucinates less in general? | ||