Abstract
Model collaboration and Switch Generation improve language model performance by leveraging the strengths of different models in a dynamic and complementary manner.
Alignment training has tradeoffs: it helps language models (LMs) gain in reasoning and instruction following but might lose out on skills such as creativity and calibration, where unaligned base models are better at. We aim to make the best of both worlds through model collaboration, where different models in the training pipeline collaborate and complement each other. Since LM responses feature interleaving skills that favor different models, we propose Switch Generation, where pretrained and aligned model versions take turns to ``speak'' in a response sequence. Specifically, we train a switcher LM by learning from outcomes of choosing different models to generate the next segment across diverse queries and contexts. At inference time, the switcher LM guides different model checkpoints to dynamically generate the next segment where their strengths are most needed. Extensive experiments with 8 model collaboration baselines and 18 datasets show that 1) model collaboration consistently outperforms individual models on 16 out of 18 tasks, and 2) Switch Generation further outperforms baselines by 12.9% on average. Further analysis reveals that Switch Generation discovers compositional skills to solve problems where individual models struggle and generalizes to unseen models and tasks, reusing and repurposing by-products in expensive model training pipelines that are otherwise discarded.
Community
Don't Throw Away Your Pretrained Model
This is an automated message from the Librarian Bot. I found the following papers similar to this paper.
The following papers were recommended by the Semantic Scholar API
- Enhancing Large Language Model Reasoning via Selective Critical Token Fine-Tuning (2025)
- IA2: Alignment with ICL Activations Improves Supervised Fine-Tuning (2025)
- Learning to Refine: Self-Refinement of Parallel Reasoning in LLMs (2025)
- PIKA: Expert-Level Synthetic Datasets for Post-Training Alignment from Scratch (2025)
- TORSO: Template-Oriented Reasoning Towards General Tasks (2025)
- Jointly Reinforcing Diversity and Quality in Language Model Generations (2025)
- ToolRM: Outcome Reward Models for Tool-Calling Large Language Models (2025)
Please give a thumbs up to this comment if you found it helpful!
If you want recommendations for any Paper on Hugging Face checkout this Space
You can directly ask Librarian Bot for paper recommendations by tagging it in a comment:
@librarian-bot
recommend
Models citing this paper 2
Datasets citing this paper 0
No dataset linking this paper
Spaces citing this paper 0
No Space linking this paper
Collections including this paper 0
No Collection including this paper