Aligning LLMs with Biomedical Knowledge using Balanced Fine-Tuning

arXiv:2511.21075 · cs.LG, cs.AI · Submitted 2025-11-26 · Read on arXiv

cs.LG, cs.AI

Submitted: 2025-11-26

Updated: 2026-10-06

Comments: Accepted at the 40th Conference on Neural Information Processing Systems (NeurIPS 2026). Related work updated

Code: https://github.com/TencentAILabHealthcare/BFT

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Terminology

• Balanced Fine-Tuning — 2× in this paper
• Fine-Tuning — 2× in this paper · explained in 5 episodes
• Alignment — 1× in this paper · explained in 1 episode
• Downstream Tasks — 1× in this paper · explained in 2 episodes
• Epistemic Uncertainty — 1× in this paper · explained in 10 episodes
• GRPO — 1× in this paper · explained in 6 episodes

Sources

• On the Generalization of SFT: A Reinforcement Learning Perspective with Reward Rectification
• Large Language Models as Biomedical Hypothesis Generators: A Comprehensive Evaluation
• VCWorld: A Biological World Model for Virtual Cell Simulation
• DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models
• Rethinking On-Policy Distillation of Large Language Models: Phenomenology, Mechanism, and Recipe
• Self-Distilled Reasoner: On-Policy Self-Distillation for Large Language Models
• Self-Distillation Enables Continual Learning
• A Survey of Reinforcement Learning for Large Reasoning Models
• Reinforcement Learning Meets Large Language Models: A Survey of Advancements and Applications Across the LLM Lifecycle
• Supervised Fine Tuning on Curated Data is Reinforcement Learning (and can be improved)
• UFT: Unifying Supervised and Reinforcement Fine-Tuning
• Anchored Supervised Fine-Tuning
• ProFit: Leveraging High-Value Signals in SFT via Probability-Guided Token Selection
• RAFT: Reward rAnked FineTuning for Generative Foundation Model Alignment
• Contract Memory Compiler: Resolve, Then Traverse
• UnityShots: Memory-Driven Multi-Shot Audio-Video Generation with Boundary-Aware Gating
• Text and Code Embeddings by Contrastive Pre-Training
• When Images Look Right and Retrieve Wrong: Coverage-Guided Cross-Scale Re-Indexing for Knowledge-Faithful Generative Perception

Related papers

• Polynomial-Augmented Neural Networks (PANNs) with Weak Orthogonality Constraints for Enhanced Function and PDE Approximation
• AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning
• Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation
• Convergence issues in Relational Concept Analysis based on AOC-posets
• Beliefs Beyond Posteriors: The Predictive Cost of ELBO Binding
• Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks