Chain-of-thought (CoT) reasoning in large language models often produces redundant steps that increase generation cost and obscure when reasoning has effectively completed. Existing stopping methods rely on output-level or token-level signals, overlooking internal representation dynamics that reflect reasoning progress. We propose STARS (STate-Aware Reasoning Stopping), a framework that leverages spectral analysis of hidden-state representations to detect transitions in reasoning phases. Our key insight is that structural changes in the representation space capture shifts in reasoning states, enabling early and reliable stopping decisions. To address representation instability, where semantically similar steps can induce different hidden states, we focus on spectral structure rather than raw activations, improving robustness to context-dependent drift. Experiments on four mathematical and logical reasoning benchmarks show that STARS reduces generation cost by over 45% under fixed token budgets while improving accuracy over standard CoT baselines.

Agentic language models solve tasks through reasoning, tool calls, observations, and terminal answers, exposing rich process information beyond the final answer. For practitioners, knowing which process patterns lead to success or failure is essential to diagnose agent errors and build better process reward models without relying on final-answer supervision. However, existing analyses often evaluate only final outcomes or, when examining traces, flatten them into homogeneous step sequences, fragmenting each reasoning round into disconnected steps. Our key insight is that an agent’s reasoning, action, and observation within a round are better treated as typed segments of a single underlying process state, rather than independent steps. We propose MULTI-VIEW-RA, a reasoning abstraction that represents each round through two complementary views: accumulated state and within-round change. Using unlabeled clustering and binary outcome labels, we apply MULTI-VIEW-RA to two practical use cases: diagnosing failures by surfacing the process states that separate correct from incorrect trajectories, and providing a transition-level correctness signal that is a candidate source of process supervision. Across Math, HotpotQA, and FEVER with multiple Qwen and Llama models, MULTI-VIEW-RA outperforms step-level explainability baselines in cases with sufficient success rates, with particularly strong gains on longer, tool-rich traces.

Large Language Models (LLMs) are susceptible to indirect prompt injection attack, where the model inadvertently responds to instructions injected into the prompt context. This vulnerability stems from LLMs’ inability to distinguish between data and instructions within a prompt. We propose CachePrune that defends against this attack by identifying and pruning neurons associated with instruction-following, during KV cache encoding of the prompt context. The pruning steers the LLM toward interpreting the context purely as data rather than as instructions to follow. To identify these neurons, we introduce a neural attribution mechanism guided by a preferential attribution loss, and theoretically connect this loss to an upper bound of the Direct Preference Optimization (DPO) objective. Further, we improve on the fidelity of neural attribution by leveraging an observed triggering effect in instruction-following. Our approach does not interfere with prompt formatting or incur test-time overhead in response generation. Experiments show that CachePrune significantly reduces the attack success rate while preserving the LLM’s ability to follow user instructions.

Instruction tuning in multimodal large language models (MLLMs) generally involves cooperative learning between a backbone LLM and a feature encoder of non-text input modalities. The major challenge is how to efficiently find the synergy between the two modules so that LLMs can adapt their reasoning abilities to downstream tasks while feature encoders can adjust to provide more task-specific information about its modality. In this paper, we analyze the MLLM instruction tuning from both theoretical and empirical perspectives, where we find the unbalanced learning between the feature encoder and the LLM can cause problems of oscillation and biased learning that lead to sub-optimal convergence. Inspired by our findings, we propose a Multimodal Balance Coefficient that enables quantitative measurement of the balance of learning. Based on this, we further design a dynamic learning scheduler that better coordinates the learning between the LLM and feature encoder, alleviating the problems of oscillation and biased learning. In addition, we introduce an auxiliary regularization on the gradient to promote updating with larger step sizes, which potentially allows for a more accurate estimation of the proposed MultiModal Balance Coefficient and further improves the training sufficiency. Our proposed approach is agnostic to the architecture of LLM and feature encoder, so it can be generically integrated with various MLLMs. We conduct experiments on multiple downstream tasks with various MLLMs, demonstrating that the proposed method is more effective than the baselines in MLLM instruction tuning.

Previous work of class-incremental learning for Named Entity Recognition (NER) relies on the assumption that there exists abundance of labeled data for the training of new classes. In this work, we study a more challenging but practical problem, i.e., few-shot class-incremental learning for NER, where an NER model is trained with only few labeled samples of the new classes, without forgetting knowledge of the old ones. To alleviate the problem of catastrophic forgetting in few-shot class-incremental learning, we reconstruct synthetic training data of the old classes using the trained NER model, augmenting the training of new classes. We further develop a framework that distills from the existing model with both synthetic data, and real data from the current training set. Experimental results show that our approach achieves significant improvements over existing baselines.