The manufacturing sector is increasingly adopting Multimodal Large Language Models (MLLMs) to move from low-level perception toward higher-level reasoning and decision support, yet current evaluations rarely reflect the fine-grained precision that real-world manufacturing demands. Progress is further hindered by data scarcity and a lack of fine-grained domain semantics in existing datasets. To bridge this gap, we introduce FORGE, a benchmark for three core quality-assurance tasks of discrete manufacturing: workpiece verification, structural surface inspection, and assembly verification. We first construct a high-quality multimodal dataset spanning real-world 2D images and 3D point clouds, annotated with fine-grained domain semantics such as standardized model numbers. We then evaluate 18 state-of-the-art MLLMs on these three tasks, revealing significant performance gaps. Counter to conventional understanding, our bottleneck analysis on five representative models indicates that visual grounding is not the primary limiting factor. Instead, insufficient domain-specific knowledge is the key bottleneck. Beyond evaluation, we show that our structured annotations can serve as an actionable training resource: supervised fine-tuning of a compact 3B-parameter model on our data yields up to 90.8% relative improvement in accuracy on held-out manufacturing scenarios, providing preliminary evidence for a practical pathway toward domain-adapted manufacturing MLLMs.

Test-Time Scaling has emerged as a powerful method to extend the reasoning capabilities of Large Language Models. However, single-agent TTS faces significant scalability bottlenecks, as excessively long reasoning traces lead to increased inference costs and stability issues caused by context management failures. To address these limitations, we propose leveraging Multi-Agent Systems as a structural upgrade to standard TTS. By decomposing monolithic reasoning chains into distinct, manageable contexts across multiple agents, MAS offers a more robust framework for scaling reasoning. We validate this approach by introducing M500, a dataset comprising 500 high-quality multi-agent, multi-turn collaborative reasoning traces generated via DeepSeek-R1. Through Supervised Fine-Tuning on M500, we enable open-source models to internalize collaborative reasoning patterns and show improved TTS performance in MAS. Furthermore, we propose an adaptive scaling strategy incorporating a “CEO” agent to dynamically guide the reasoning process and optimize collaboration depth. Extensive experiments within the AgentVerse framework demonstrate that our fine-tuned models, Qwen2.5-32B-MAS and Phi4-14B-MAS, significantly outperform their base counterparts.

Large language models (LLMs) increasingly solve complex reasoning tasks via long chain-of-thought, but their forward-only autoregressive generation process is fragile; early token errors can cascade, which creates a clear need for self-reflection mechanisms. However, existing self-reflection either performs revisions over full drafts or learns self-correction via expensive training, both fundamentally reactive and inefficient. To address this, we propose Self-Reflective Generation at Test Time (SRGen), a lightweight test-time framework that reflects before generating at uncertain points. During token generation, SRGen utilizes dynamic entropy thresholding to identify high-uncertainty tokens. For each identified token, it trains a specific corrective vector, which fully exploits the already generated context for a self-reflective generation to correct the token probability distribution. By retrospectively analyzing the partial output, this self-reflection enables more trustworthy decisions, thereby significantly reducing the probability of errors at highly uncertain points. Evaluated on challenging mathematical reasoning benchmarks and a diverse set of LLMs, SRGen can significantly strengthen model reasoning. Moreover, our findings position SRGen as a plug-and-play method that integrates reflection into the generation process for reliable LLM reasoning, achieving consistent gains with bounded overhead and can be combined with other training-time (e.g., RLHF) and test-time (e.g., SLOT) techniques.

Ensuring that Large Language Models (LLMs) adhere to safety principles without refusing benign requests remains a significant challenge. While OpenAI introduces deliberative alignment (DA) to enhance the safety of its o-series models through reasoning over detailed “code-like” safety rules, the effectiveness of this approach in open-source LLMs, which typically lack advanced reasoning capabilities, is understudied. In this work, we systematically evaluate the impact of explicitly specifying extensive safety codes versus demonstrating them through illustrative cases. We find that referencing explicit codes inconsistently improves harmlessness and systematically degrades helpfulness, whereas training on case-augmented simple codes yields more robust and generalized safety behaviors. By guiding LLMs with case-augmented reasoning instead of extensive code-like safety rules, we avoid rigid adherence to narrowly enumerated rules and enable broader adaptability. Building on these insights, we propose CADA, a case-augmented deliberative alignment method for LLMs utilizing reinforcement learning on self-generated safety reasoning chains. CADA effectively enhances harmlessness, improves robustness against attacks, and reduces over-refusal while preserving utility across diverse benchmarks, offering a practical alternative to rule-only DA for improving safety while maintaining helpfulness.