Paper 1
■書誌情報
Junyu Liu, Zirui Li, Qian Niu, Zequn Zhang, Yue Xun, Wenlong Hou, Shujun Wang, Yusuke Iwasawa, Yutaka Matsuo, Kan Hatakeyama-Sato: JMedEthicBench: A Multi-Turn Adversarial Benchmark for Japanese Medical Ethics Alignment in LLMs, Proceedings of the 3rd Conference on Language Modeling (COLM 2026), October 2026
■概要
As Large Language Models (LLMs) are increasingly deployed in healthcare worldwide, robustness against adversarial jailbreaking becomes essential for maintaining medical ethics compliance, calling for comprehensive adversarial benchmarks. However, existing benchmarks remain predominantly English-centric and limited to single-turn attacks. To address these gaps, we introduce JMedEthicBench, the first multi-turn adversarial benchmark for evaluating medical ethics alignment of LLMs in Japanese. Grounded in 67 guidelines from the Japan Medical Association, our benchmark comprises over 50,000 adversarial conversations generated using seven automatically discovered jailbreak strategies. Through a dual-LLM scoring protocol, we evaluate 22 models and find that commercial models maintain strong resistance to attacks while medical-specialized models exhibit increased vulnerability. Furthermore, safety scores decline significantly across conversation turns (median: 9.5 to 5.5, ), demonstrating that multi-turn escalation systematically circumvents defense mechanisms. Cross-lingual evaluation reveals that medical model vulnerabilities persist across Japanese and English, indicating inherent alignment limitations rather than language-specific factors. Our findings suggest that domain-specific fine-tuning may inadvertently weaken safety alignment and that multi-turn jailbreak attacks represent a distinct threat surface requiring dedicated defense strategies.
Paper 2
■書誌情報
Yongmin Kim, Shota Takashiro, Takeshi Kojima, Yusuke Iwasawa, Yutaka Matsuo: Batch-wise Adaptive Pruning: Periodic Neuron Activation-Aware Weight Pruning for Language Reasoning Model, Proceedings of the 3rd Conference on Language Modeling (COLM 2026), October 2026
■概要
Large Reasoning Models (LRMs) achieve strong performance on complex tasks through extended chain-of-thought generation, but incur substantial computational costs during inference. In production settings, batched inference is essential for high throughput, yet existing adaptive pruning methods face performance limitations: First, they rely on averaging activations across samples to determine shared pruning masks, which may miss critically activated neurons for some individual samples. Second, after the averaging, they adopt threshold-based selection for pruning neurons, causing sparsity ratio instability. In this work, we propose a training-free adaptive pruning method designed specifically for batched inference in LRMs. Since averaging activations across samples can miss neurons that are critical for individual samples, our method adopts max-pooling for cross-sample aggregation to preserve such sample-specific important neurons. To stabilize the activation sparsity ratio, our method adopts periodic top-k selection over the aggregated neurons instead of threshold-based selection. Furthermore, based on the observation that important neurons tend to be repeatedly activated, we incorporate an activation memory mechanism to capture periodically important neurons. Experiments on diverse reasoning benchmarks demonstrate that our method outperforms the previous state-of-the-art adaptive pruning method by 39.7 percentage points in average accuracy at batch size 4 with 50% target sparsity on DeepSeek-R1-Distill-Qwen-7B, along with 1.40× speedup over dense inference at 50% actual sparsity, demonstrating practical efficiency gains for deployment.
