Limited controlled evidence exists on how training data, adaptation method, and model scale jointly affect tool-calling performance in language-model agents. We evaluate supervised fine-tuning (SFT) with LoRA, reinforcement learning (RL) via Group Relative Policy Optimization (GRPO), and SFT followed by GRPO across six Qwen3 models from 0.6B to 32B parameters, covering both in-distribution performance and cross-dataset transfer. SFT with LoRA is the strongest in-distribution method throughout the 0.6B-32B range and best in 15 out of 18 experimental settings. On cross-dataset transfer, the methods are closer: GRPO wins 29 out of 54 settings where training and test datasets differ, but its margin over SFT averages under one point, and SFT->GRPO is rarely strongest in either comparison. Dataset mixing gives consistently strong transfer while staying close to specialized in-distribution training, regardless of method. Additional analysis further confirms that LoRA outperforms full-parameter fine-tuning, demonstrating that LoRA better preserves pretrained agentic behavior.
DataFlex-RL, an evaluation platform for comparing choices under a common GRPO recipe, is introduced, finding that changing the data policy measurably changes the training process but does not produce a reproducible improvement over uniform training.
Hao Liang, Ming-Rui Chen, Hengyi Feng et al.· 2 citations
Group Relative Policy Optimization (GRPO) has emerged as a memory-efficient reinforcement fine-tuning (RFT) technique for reasoning-intensive tasks. How- ever, GRPO training dynamics on small language models (SLMs) remain poorly understood, limiting its reliable adoption and reproducibility in open and resource- constr...
Rajat Ghosh, Vaishnavi Bhargava, Henry Wong et al.· 0 citations
Supervised fine-tuning (SFT) on offline agent trajectories is the standard approach for training specialized tool-using agents, but forcing models to imitate reasoning and actions token by token may harm other capabilities (e.g., general reasoning, tool calling, code generation) of the base model. In this work, we focu...
Rong-Hua Li, Zi Liang, Zhi-Shan Li et al.· 0 citations
This work studies how to balance SFT and RL under production-mirroring beta APIs, and observes that, in a controlled experiment, checkpoint trajectories retrospectively separated into three regimes: Imitation, where SFT captured reliable teacher behavior; Lift, where both stages helped; and Discovery, where useful rewa...
Aakash Kolekar, Sahika Genc, Bunyamin Sisman et al.· 0 citations
Reinforcement learning (RL) has become a central post-training approach for reasoning and agentic large language models (LLMs), particularly when task outcomes can be verified automatically. Comparisons across this literature remain difficult because a reported gain may combine changes to the learning signal, policy co...
Liu Yang, Han Zhu, Zheng-Yang Zhong et al.· Unmanned Systems· 1 citation
Critic-free reinforcement fine-tuning (RFT) for agentic large language models is often done through GRPO-style methods, which compute a group baseline over repeated rollouts to reduce target variance. However, this setup is ill-suited to agents acting in stateful environments such as live services or security sandboxes...
Joery A. de Vries, Neil D. Lawrence, Zhen-Wen Dai· 0 citations
Related blog posts
MIT News · Artificial Intelligence· news.mit.eduSep 24, 2026
A new method, called CW-Net, translates the reasoning process of an autonomous vehicle’s AI system into understandable concepts that explain its behavior.
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
