Fine-tuning LLMs for agent tasks involves training models on domain-specific data to improve their reliability at tool calling, instruction following, and structured reasoning. While prompt engineering and RAG handle many use cases, fine-tuning becomes essential when agents need consistent behavior on specialized tasks, structured output compliance, or optimized performance at reduced model sizes and costs.

Rule of thumb: Start with prompt engineering. If evaluation shows consistent failures on specific behaviors after prompt optimization, fine-tune.

Train on curated (prompt, completion) pairs that demonstrate desired agent behavior. For tool-use agents, this includes examples of correct function calls, argument formatting, and multi-step reasoning chains.

LoRA (Low-Rank Adaptation) inserts small trainable matrices into frozen model layers, reducing compute by 10-100x while maintaining performance. QLoRA (Quantized LoRA) adds 4-bit quantization, enabling billion-parameter model fine-tuning on consumer GPUs.

from peft import LoraConfig, get_peft_model, TaskType
from transformers import AutoModelForCausalLM, AutoTokenizer, TrainingArguments
from trl import SFTTrainer
 
# Load base model
model = AutoModelForCausalLM.from_pretrained(
    "meta-llama/Llama-3-8B-Instruct",
    load_in_4bit=True  # QLoRA quantization
)
tokenizer = AutoTokenizer.from_pretrained("meta-llama/Llama-3-8B-Instruct")
 
# Configure LoRA adapters
lora_config = LoraConfig(
    task_type=TaskType.CAUSAL_LM,
    r=16,               # Rank of low-rank matrices
    lora_alpha=32,       # Scaling factor
    lora_dropout=0.05,
    target_modules=["q_proj", "v_proj", "k_proj", "o_proj"]
)
model = get_peft_model(model, lora_config)
 
# Train on tool-calling dataset
trainer = SFTTrainer(
    model=model,
    tokenizer=tokenizer,
    train_dataset=tool_calling_dataset,  # (prompt, tool_call) pairs
    args=TrainingArguments(
        output_dir="./agent-lora",
        num_train_epochs=3,
        per_device_train_batch_size=4,
        learning_rate=2e-4,
        warmup_steps=100
    )
)
trainer.train()

Aligns agent behavior with human preferences through three phases:

1. Collect comparisons — Humans rank agent outputs for the same input
2. Train reward model — A model learns to score outputs based on human preferences
3. Optimize with PPO — The agent is trained via reinforcement learning to maximize the reward model's score

RLHF produces safer, more helpful agents but requires significant human annotation effort.

Simplifies RLHF by directly optimizing on preference pairs without training a separate reward model. DPO is more stable and computationally efficient, making it practical for smaller teams fine-tuning agent behavior.

Effective fine-tuning for function calling requires curated datasets:

• Function call pairs — (user_query, correct_tool_call_with_arguments) examples demonstrating proper invocation
• Multi-step traces — Complete agent trajectories showing planning, tool calls, and synthesis
• Error recovery examples — Demonstrations of handling failed tool calls gracefully
• Negative examples — Cases where no tool should be called, teaching the model restraint

Public datasets include Gorilla APIBench for API calling and xLAM Function Calling for structured tool use.

• Loss convergence — Monitor training and validation loss for overfitting
• Function calling accuracy — Percentage of correct tool selections and argument formatting
• BFCL benchmark — Berkeley Function Calling Leaderboard scores before and after fine-tuning
• Task completion rate — End-to-end success on representative agent tasks
• Regression testing — Ensure fine-tuning doesn't degrade general capabilities

• SimpliSmart - Fine-Tuning LLMs in 2025
• Towards AI - Fine-Tuning Techniques and Trade-offs
• Gorilla - LLM API Calling Benchmark

• Function Calling — The tool-calling capability that fine-tuning improves
• Embeddings — Fine-tuning embedding models for better retrieval
• Prompt Engineering — Prompt optimization as an alternative to fine-tuning
• Agent Debugging — Evaluating fine-tuned agent performance