The paper introduces ContinuousBench, a dynamic benchmark designed to rigorously test if differentially private (DP) synthetic text can genuinely transfer new knowledge and capabilities from sensitive…

The paper introduces ContinuousBench, a novel benchmark designed to rigorously test if differentially private (DP) synthetic text can genuinely transfer new knowledge, finding that state-of-the-art DP…

The paper investigates the limits of self-evolution in LLM reasoning under closed-loop settings, finding that while self-improvement is significant, it consistently falls short of perfect oracle super…

The paper introduces a novel, transferable learned attack (LT-MIA) that detects a universal 'signature of memorization' in language models, achieving high accuracy across diverse model architectures (…

This paper introduces Capability Self-Assessment (CSA), a crucial ability for LLMs to recognize their limitations, and demonstrates that reinforcement learning is an effective method for teaching this…

This paper systematically investigates unlearnable examples (UEs) across diverse training paradigms, finding that existing UEs fail under pretraining-finetuning (PF) settings, and proposes Shallow Sem…

The paper demonstrates that the AI-like style introduced by post-training alignment can be measured, localized, and causally removed using a novel ablation technique called PASTA.

The paper proposes Distribution-Aligned Self-Distillation (DASD) to improve self-distillation by dynamically filtering high-perplexity tokens, thereby preserving useful logical knowledge while suppres…

Divergence Decoding (DD) is a novel, effective, and inexpensive method that uses auxiliary models to steer LLM logits during inference, enabling the removal of memorized sensitive data without signifi…

The paper introduces Responsible Contrastive Soft Prompting (RCSP), a parameter-efficient method using soft prompts to improve LLM reliability by simultaneously suppressing hallucinations, encouraging…

The paper introduces StreamSynth, a sequential setting for synthetic data generation, and proposes SynLearner, a framework that enables LLMs to improve synthesis performance by accumulating and transf…

The paper introduces Canonical-Context On-Policy Distillation (CCOPD) to improve multi-turn language model performance by mitigating 'self-anchored drift,' ensuring consistent answers regardless of wh…

The paper proposes COSE, a method that uses an LLM's intrinsic confidence as an uncertainty signal to improve self-evolutionary training, achieving state-of-the-art performance on general reasoning an…

The paper proposes using fine-grained quality signals, such as pairwise self-judgments and token-level entropy, instead of simple binary correctness to improve LLM performance on saturated datasets, s…

The paper introduces SARDI, a novel, training-free framework that uses low-confidence 'lookahead' tokens generated during the denoising process of discrete diffusion language models to dynamically gui…

The paper proposes a debiasing fine-tuning technique to efficiently enhance the robustness of Large Language Models against semantically similar but textually altered prompts.

The paper introduces DOMINO, a novel inductive framework that synthesizes domain-specific data for LLMs using only reference examples, significantly improving performance on challenging, implicitly de…

This paper introduces a new scaling law for sparse language models trained with limited data, demonstrating that sparsity can significantly improve performance and delay data saturation during multi-e…

This paper proposes Supervised Memory Training (SMT), a method for training nonlinear RNNs that sidesteps recurrent credit propagation entirely.