Alyssa L

AGENTRUN: Failure-Driven QLoRA Specialization for Multi-Tool Agent Behavior in a 7B Language Model

Parameter-efficient fine-tuning is widely assumed to improve small open-weight models on the specific behaviors it targets. We present a controlled case study that complicates this assumption for multi-tool agent execution. Using Qwen2.5-7B-Instruct and a fixed 23-task, machine-c...

AGENTRUN: Failure-Driven QLoRA Specialization for Multi-Tool Agent Behavior in a 7B Language Model

Parameter-efficient fine-tuning is widely assumed to improve small open-weight models on the specific behaviors it targets. We present a controlled case study that complicates this assumption for multi-tool agent execution. Using Qwen2.5-7B-Instruct and a fixed 23-task, machine-c...