Open-domain event detection aims to identify known event types while discovering previously unseen ones from unstructured documents. In practice, event detection systems often operate with an evolving event set rather than a complete type inventory. Under this setting, existing zero-shot event detection methods can fail in two ways when the event type is absent from the predefined set: the model may either miss the event or assign it to a plausible but wrong known type, hiding the need for discovery. To address this problem, we propose OGED, an  ̲Ontology- ̲Guided open-domain  ̲Event  ̲Detection framework for grounding–discovery resolution, i.e., deciding whether an event mention should be assigned to an existing type or should induce a new type. OGED organizes the available event types into a hierarchical search structure, uses cluster-level scheduler and adaptive planning to keep known-type assignment and new-type discovery jointly active, and applies retrieval-gated verification together with conflict selection to resolve competing assignments before output. Experiments on a unified benchmark with 564 event types show that OGED outperforms strong baselines on both known-type detection and unknown-type discovery. Code is available at https://github.com/chenzhouli/OGED.

Safety-aligned large language models (LLMs) often generate refusal responses to harmless queries due to the over-refusal problem. However, existing methods for mitigating over-refusal cannot maintain a low refusal ratio for harmless queries while keeping a high refusal ratio for malicious ones. In this paper, we analyze how system prompts with varying safety levels affect LLM refusal behaviors when facing over-refusal queries. A key observation is that, when LLMs suffer from the over-refusal issue, non-refusal tokens remain present in the next-token candidate list, but the model systematically fails to select them, despite the generation of refusal tokens. Based on this observation, we propose a training-free and model-agnostic approach, Adaptive Contrastive Decoding (AdaCD), to mitigate over-refusal while maintaining LLM safety. First, AdaCD compares the output distributions of the LLM with or without an extreme safety system prompt to refine the refusal token distribution. Second, we introduce an adaptive contrastive decoding strategy that dynamically incorporates or removes the refusal token distribution, adaptively boosting the probability of selecting refusal or non-refusal tokens. Experimental results on five benchmark datasets show that, on average, AdaCD reduces the refusal ratio for over-refusal queries by 10.35%, yet still increases the refusal ratio for malicious queries by 0.13%. Code is available at https://shorturl.at/Z31Oe.

As large language models (LLMs) are increasingly applied across various domains, enhancing safety while maintaining the helpfulness of LLMs has become a critical challenge. Recent studies solve this problem through safety-constrained online preference optimization or safety-constrained offline preference optimization. However, the safety-constrained online methods often suffer from excessive safety, which might reduce helpfulness, while the safety-constrained offline methods perform poorly in adaptively balancing safety and helpfulness. To address these limitations, we propose MidPO, a Mixture of Experts (MoE) framework for safety-helpfulness dual Preference Optimization. Firstly, MidPO devises single-preference enhanced direct preference optimization approach to transform the base model into two independent experts, termed safety and helpfulness experts, and fine-tunes the two independent experts for optimal safety or helpfulness performance. Secondly, to achieve an effective balance between safety and helpfulness, MidPO incorporates the two experts into the MoE framework and designs a dynamic routing mechanism to allocate contributions from each expert adaptively. We conduct quantitative and qualitative experiments on three popular datasets to demonstrate the proposed MidPO significantly outperforms state-of-the-art approaches in both safety and helpfulness. Code is available at https://github.com/OutdoorManofML/MidPO.