Instruction-Tuned, but Not More Verifiable Instruction-Following: A Cross-Task Diagnosis for LoRA Adapters

1. 1Evaluate LoRA adapters across multiple tasks to assess capability gains
2. 2Measure instruction-following capabilities using automatically verifiable metrics like IFEval
3. 3Compare nominal training objectives with realized cross-task capability gains
4. 4Analyze results across multiple seeds to ensure consistency

ML researchers and engineers benefit from understanding the limitations of instruction-tuned models, as it impacts the development and deployment of reliable AI systems

💡 Nominal training objectives may not align with realized capability gains, highlighting the need for rigorous evaluation

🤖 Instruction-tuned LoRA adapters may not improve instruction-following capabilities as expected #LLMs #AI