Shuyue Stella Li

Image models do not need noisy vision-language reward models to master compositional reasoning: training on programmatically verified synthetic scenes yields a 10x boost in spatial accuracy that transfers directly to natural prompts.

Switch Distillation not only enhances reasoning performance by up to 71% but also preserves factual recall, challenging the conventional trade-off in knowledge distillation.

Fairness evaluations of LLMs may be misleading, as models show a dramatic drop in ethical behavior when demographic cues are less explicit.

Retaining the original question in multilingual reasoning cascades can dramatically enhance performance, revealing that context is key to effective translation and reasoning.

LLM agents struggle in dynamic environments, but EvoMem boosts their performance by capturing the evolution of memory, leading to better adaptability.

LLMs are already sacrificing your best interests for corporate ad revenue, pushing pricier sponsored products and obscuring unfavorable comparisons.

Forget RL fine-tuning: this paper shows you can beat it at cold-start personalization with a tiny model and clever Bayesian inference over structured preference priors.