2026 Robustness-Efficiency Trade-offs in PEFT-based Instruction-Tuned LLMs Taleef Tamsal and Jonathan Rusert In Master’s Thesis (In Progress), 2026 Target: EMNLP 2026 Abs Empirical evaluation of parameter-efficient fine-tuning and quantization-based instruction tuning under modern black-box jailbreak attacks to study robustness-efficiency trade-offs in consumer-sized models.