Dataset · lambda
Multi-turn tool-calling reasoning traces for agents
lambda/hermes-agent-reasoning-traces
Real agent dialogues with step-by-step reasoning and tool execution results in English.
What's inside
Contains multi-turn tool-calling trajectories for training AI agents in the Hermes Agent harness. Each example is a real agent conversation with <think> reasoning blocks and actual tool execution results. Two configs: 7,646 examples from Kimi-K2.5 and 7,055 from GLM-5.1-FP8; language is English.
What you can do with it
- Agent SFT for multi-turn tool-calling
- LoRA training on Hermes Agent trajectories
- Model evaluation on tool invocation
- Tasks
- Text generation
- Languages
- English
- Size
- 10K–100K rows
- Downloads this month
- 2.6K
- Likes
- 396
Models trained on it
- nicosuter/Qwen3.8-27B-AWQImage text → text 436.5K
- ManniX-ITA/gemma-4-26B-A4B-agentic-GGUFText generation 15.8K
- Jackrong/Qwopus3.6-27B-Coder-MTP-GGUFImage text → text 11.8K
- mradermacher/gemma-4-26B-A4B-agentic-i1-GGUF 9.2K
- Jackrong/Qwopus3.6-27B-Coder-Compat-MTP-GGUFImage text → text 8.3K
- Jackrong/Qwopus3.5-9B-Coder-GGUFImage text → text 7.9K
- mradermacher/Qwen3.5-4B-Claude-3xMerge-Jackrong-UNCENSORED-heretic-i1-GGUF 7.5K
- Jackrong/Qwopus3.5-4B-Coder-MTP-GGUFImage text → text 6.9K
How to load it
from datasets import load_dataset
ds = load_dataset("lambda/hermes-agent-reasoning-traces")