1488 successful agent rollouts from the moonshot/kimi-k2.5 model on the seta-env-v2 terminal-agent benchmark were distilled into this supervised fine-tuning dataset. The dataset was created by camel-ai and tokenized with the Qwen/Qwen3-8B chat template, ready for training. It was last updated on April 7, 2026.
Use Cases
- Supervised fine-tuning of language models for agent behavior based on successful terminal-agent rollouts.
- Inspecting and filtering diagnostic records from agent trials for analysis.
- Re-tokenizing training data for different model architectures using the preserved per-trial records.
Strengths
- Contains 1488 successful agent rollouts, providing a substantial set of positive examples.
- Preserves full per-trial diagnostic records, enabling inspection and re-processing without rerunning simulations.
Limitations
- Column-level documentation is absent; field semantics must be inferred after download.
- Row count is unknown, which may limit suitability assessment.
Provenance
- Source
- camel-ai
- Collection Method
- Distilled from successful agent rollouts of the moonshot/kimi-k2.5 model on the seta-env-v2 terminal-agent benchmark.
- Time Range
- null
- Freshness
- Last updated 2026-04-07 19:36:59; freshness should be verified.
- Geography
- null