Public KG record
ALTK-Evolve
Recorded overview
Turns past trajectories into individually retrievable guidelines with support co
An evidence-backed entity from the project knowledge graph; detailed tool metadata is still being added.
Recorded details
- Benchmarks: τ²-bench retail：Pass^4 为 67.5%
- Benchmarks: 零售 Mean Accuracy 从 78.1% 升至 82.5%
- Benchmarks: 航空 Mean Accuracy 从 57.5% 升至 65.0%
- Benchmarks: 航空 Pass^4 从 35.0% 升至 45.0%
- Benchmarks: AppWorld
- Capabilities: 从代理历史轨迹生成可复用指南
- Capabilities: 在推理时注入生成的指南
- Capabilities: consistency guidelines
Evidence coverage
The KG currently records 9 known information dimensions and has 10 undocumented dimensions, linked to 5 public source articles. These counts describe record coverage, not model performance or verified fit.
Related public records
- Arbitrary code execution — alternative to
- ReAct — based on
- Arbitrary code execution — competes with
- guideline set — develops
- AppWorld — evaluated on