language model training
5 notes, newest first.
How Does Alignment Tuning Shape Representations of Sycophancy and Related Cue-Induced Biases in LLMs?
The research investigates the susceptibility of large language models (LLMs) to biases induced by subtle changes in input prompts. The study finds that these biases are largely due to alignment tuning rather than…
OR Else: A Differentiable Trust Region for Policy Optimization
The paper compares PPO (Proximal Policy Optimization) and GRPO (Group Regularized Policy Optimization) with two variations using a differentiable trust region for policy optimization called Output Reset (OR). The study…
CausalMix: Data Mixture as Causal Inference for Language Model Training
The paper presents CausalMix, a method that addresses the limitation of existing data mixing methods in Large Language Model (LLM) training by casting it as a causal inference problem. By using causal modeling to…
Right in the Right Way: LM Training with Verifiable Rewards and Human Demonstrations
Researchers propose an adversarial generator-discriminator framework to train language models (LM) for tasks with well-defined success metrics, like code generation and mathematical reasoning. The new approach improves…
OpenThoughts-Agent: Data Recipes for Agentic Models
The OpenThoughts-Agent project presents a fully open data curation pipeline for training agentic models, which are language models that expand AI applications. The project conducted controlled ablation experiments and…
