A paper published on October 6, 2025 proposes the MoE-CL framework for continual instruction tuning and reports evaluations on the MTL5 and Tencent3 benchmarks, plus an A/B test on Tencent. The available summary gives no quantitative results.
Published on October 6, 2025, the paper proposes MoE-CL, a parameter-efficient mixture-of-experts framework for continual instruction tuning. It combines task-specific and shared LoRA experts with a task-aware discriminator, aiming to address catastrophic forgetting as language models adapt to changing tasks.
The summary reports evaluations on the MTL5 and Tencent3 benchmarks and an A/B test on Tencent, but provides no detailed or quantitative findings. To assess the conclusions, consult the original paper and check its full methods, metrics, and results. If you use AI to study or apply the work, avoid submitting personal data or confidential documents without authorization, and check your organization’s policies.