Creativity assessment in science and engineering is increasingly based on both human and AI judgment, but the cognitive processes and biases behind these evaluations remain poorly understood. We conducted two experiments examining how including example solutions with ratings impact creativity evaluation, using a finegrained annotation protocol where raters were tasked with explaining their originality scores and rating for the facets of remoteness (whether the response is "far" from everyday ideas), uncommonness (whether the response is rare), and cleverness. In Study 1, we analyzed creativity ratings from 72 experts with formal science or engineering training, comparing those who received example solutions with ratings (example) to those who did not (no example). Computational text analysis revealed that, compared to experts with examples, no-example experts used more comparative language (e.g., "better/worse") and emphasized solution uncommonness, suggesting they may have relied more on memory retrieval for comparisons. In Study 2, parallel analyses with state-of-the-art LLMs revealed that models prioritized uncommonness and remoteness of ideas when rating originality, suggesting an evaluative process rooted around the semantic similarity of ideas. In the example condition, while LLM accuracy in predicting the true originality scores improved, the correlations of remoteness, uncommonness, and cleverness with originality also increased substantially - to upwards of 0.99 - suggesting a homogenization in the LLMs evaluation of the individual facets. These findings highlight important implications for how humans and AI reason about creativity and suggest diverging preferences for what different populations prioritize when rating.

Scaling inference compute enhances reasoning in large language models (LLMs), with long chains-of-thought (CoTs) enabling strategies like backtracking and error correction. Reinforcement learning (RL) has emerged as a crucial method for developing these capabilities, yet the conditions under which long CoTs emerge remain unclear, and RL training requires careful design choices. In this study, we systematically investigate the mechanics of long CoT reasoning, identifying the key factors that enable models to generate long CoT trajectories. Through extensive supervised fine-tuning (SFT) and RL experiments, we present four main findings: (1) While SFT is not strictly necessary, it simplifies training and improves efficiency; (2) Reasoning capabilities tend to emerge with increased training compute, but their development is not guaranteed, making reward shaping crucial for stabilizing CoT length growth; (3) Scaling verifiable reward signals is critical for RL. We find that leveraging noisy, web-extracted solutions with filtering mechanisms shows strong potential, particularly for out-of-distribution (OOD) tasks such as STEM reasoning; and (4) Core abilities like error correction are inherently present in base models, but incentivizing these skills effectively for complex tasks via RL demands significant compute, and measuring their emergence requires a nuanced approach. These insights provide practical guidance for optimizing training strategies to enhance long CoT reasoning in LLMs. Our code is available at: https://github.com/eddycmu/demystify-long-cot.

Diagrams play a crucial role in visually conveying complex relationships and processes within business documentation. Despite recent advances in Vision-Language Models (VLMs) for various image understanding tasks, accurately identifying and extracting the structures and relationships depicted in diagrams continues to pose significant challenges. This study addresses these challenges by proposing a text-driven approach that bypasses reliance on VLMs' visual recognition capabilities. Instead, it utilizes the editable source files--such as xlsx, pptx or docx--where diagram elements (e.g., shapes, lines, annotations) are preserved as textual metadata. In our proof-of-concept, we extracted diagram information from xlsx-based system design documents and transformed the extracted shape data into textual input for Large Language Models (LLMs). This approach allowed the LLM to analyze relationships and generate responses to business-oriented questions without the bottleneck of image-based processing. Experimental comparisons with a VLM-based method demonstrated that the proposed text-driven framework yielded more accurate answers for questions requiring detailed comprehension of diagram structures.The results obtained in this study are not limited to the tested .xlsx files but can also be extended to diagrams in other documents with source files, such as Office pptx and docx formats. These findings highlight the feasibility of circumventing VLM constraints through direct textual extraction from original source files. By enabling robust diagram understanding through LLMs, our method offers a promising path toward enhanced workflow efficiency and information analysis in real-world business scenarios.

Creating Computer-Aided Design (CAD) models requires significant expertise and effort. Text-to-CAD, which converts textual descriptions into CAD parametric sequences, is crucial in streamlining this process. Recent studies have utilized ground-truth parametric sequences, known as sequential signals, as supervision to achieve this goal. However, CAD models are inherently multimodal, comprising parametric sequences and corresponding rendered visual objects. Besides,the rendering process from parametric sequences to visual objects is many-to-one. Therefore, both sequential and visual signals are critical for effective training. In this work, we introduce CADFusion, a framework that uses Large Language Models (LLMs) as the backbone and alternates between two training stages: the sequential learning (SL) stage and the visual feedback (VF) stage. In the SL stage, we train LLMs using ground-truth parametric sequences, enabling the generation of logically coherent parametric sequences. In the VF stage, we reward parametric sequences that render into visually preferred objects and penalize those that do not, allowing LLMs to learn how rendered visual objects are perceived and evaluated. These two stages alternate throughout the training, ensuring balanced learning and preserving benefits of both signals. Experiments demonstrate that CADFusion significantly improves performance, both qualitatively and quantitatively.

Recent work has demonstrated that language models can be trained to identify the author of much shorter literary passages than has been thought feasible for traditional stylometry. We replicate these results for authorship and extend them to a new dataset measuring novel genre. We find that LLMs are able to distinguish authorship and genre, but they do so in different ways. Some models seem to rely more on memorization, while others benefit more from training to learn author/genre characteristics. We then use three methods to probe one high-performing LLM for features that define style. These include direct syntactic ablations to input text as well as two methods that look at model internals. We find that authorial style is easier to define than genre-level style and is more impacted by minor syntactic decisions and contextual word usage. However, some traits like pronoun usage and word order prove significant for defining both kinds of literary style.

This paper presents SOLOMON, a novel Neuro-inspired Large Language Model (LLM) Reasoning Network architecture that enhances the adaptability of foundation models for domain-specific applications. Through a case study in semiconductor layout design, we demonstrate how SOLOMON enables swift adaptation of general-purpose LLMs to specialized tasks by leveraging Prompt Engineering and In-Context Learning techniques. Our experiments reveal the challenges LLMs face in spatial reasoning and applying domain knowledge to practical problems. Results show that SOLOMON instances significantly outperform their baseline LLM counterparts and achieve performance comparable to state-of-the-art reasoning model, o1-preview. We discuss future research directions for developing more adaptive AI systems that can continually learn, adapt, and evolve in response to new information and changing requirements.

Content Warning: This paper contains examples of harmful language. Jailbreaking attacks can effectively manipulate open-source large language models (LLMs) to produce harmful responses. However, these attacks exhibit limited transferability, failing to disrupt proprietary LLMs consistently. To reliably identify vulnerabilities in proprietary LLMs, this work investigates the transferability of jailbreaking attacks by analysing their impact on the model's intent perception. Nevertheless, these adversarial sequences fail to mislead the target LLM's intent perception, allowing the target LLM to refocus on malicious-intent tokens and abstain from responding. Our analysis further reveals the inherent distributional dependency within the generated adversarial sequences, whose effectiveness stems from overfitting the source LLM's parameters, resulting in limited transferability to target LLMs. To this end, we propose the Perceived-importance Flatten (PiF) method, which uniformly disperses the model's focus across neutral-intent tokens in the original input, thus obscuring malicious-intent tokens without relying on overfitted adversarial sequences. Extensive experiments demonstrate that PiF provides an effective and efficient red-teaming evaluation for proprietary LLMs. Empowered by massive corpus, large language models (LLMs) have achieved human-level conversational capabilities (OpenAI, 2023a; Google, 2023; Meta, 2024) and are widely employed in real-world applications. However, their training corpus is mainly crawled from the Internet without thorough ethical review, raising concerns about the potential risks associated with LLMs. Recent red-teaming efforts highlight that jailbreaking attacks can effectively disrupt LLMs to produce undesirable content with harmful consequences (Perez et al., 2022; Ganguli et al., 2022; Ouyang et al., 2022). Unlike model-level jailbreaks that necessitate parameter modifications and are restricted to opensource LLMs (Qi et al., 2024; Huang et al., 2023a), token-level and prompt-level jailbreaks can generate transferable adversarial sequences (Yu et al., 2023; Lapid et al., 2023), thus posing a potential threat to widespread proprietary LLMs (Zou et al., 2023; Chao et al., 2023). Nevertheless, empirical results indicate that these adversarial sequences lack reliable transferability, failing to consistently manipulate target LLMs (Chao et al., 2024; Chen et al., 2024). Furthermore, these lengthy adversarial sequences can be further countered by adaptive jailbreaking detection and defence (Alon & Kamfonas, 2023; Inan et al., 2023; Robey et al., 2023; Wang et al., 2024a). As depicted in Figure 1, developing jailbreak attacks that can reliably identify vulnerabilities in proprietary LLMs--thereby promoting human alignment and preventing future misuse--remains a significant challenge. These attacks are initially generated on the source LLM (Llama-2-7B-Chat) and subsequently transferred to the target LLM (Llama-2-13B-Chat).

The introduction of 8-bit floating-point (FP8) computation units in modern AI accelerators has generated significant interest in FP8-based large language model (LLM) inference. Unlike 16-bit floating-point formats, FP8 in deep learning requires a shared scaling factor. Additionally, while E4M3 and E5M2 are well-defined at the individual value level, their scaling and accumulation methods remain unspecified and vary across hardware and software implementations. As a result, FP8 behaves more like a quantization format than a standard numeric representation. In this work, we provide the first comprehensive analysis of FP8 computation and acceleration on two AI accelerators: the NVIDIA H100 and Intel Gaudi 2. Our findings highlight that the Gaudi 2, by leveraging FP8, achieves higher throughput-to-power efficiency during LLM inference, offering valuable insights into the practical implications of FP8 adoption for datacenter-scale LLM serving.

In the training of large language models (LLMs), updating parameters more efficiently and stably has always been an important challenge. To achieve efficient parameter updates, existing methods usually achieve performance comparable to full parameter updates through methods such as low-dimensional decomposition or layer-wise selective updates. In this work, we propose AlphaAdam, an optimization framework for LLM from the perspective of intra-layer parameter updates. By decoupling parameter updates and dynamically adjusting their strength, AlphaAdam accelerates convergence and improves training stability. We construct parameter masks based on the consistency of historical momentum and gradient direction and combine them with an adaptive mask strength strategy to ensure efficient optimization and theoretical convergence guarantees, which is also applicable to most momentum-based optimizers. Extensive experiments show that AlphaAdam outperforms state-of-the-art methods such as AdamW in terms of convergence speed and computational efficiency across tasks, including GPT-2 pre-trained and fine-tuned RoBERTa and Llama-7B. Our AlphaAdam implements an optimizer enhancement framework for LLMs through intra-layer asynchronous masked adaptive updates. Our code is available in this https://github.com/MaeChd/AlphaAdam.

The rapid expansion of web content has made on-device AI assistants indispensable for helping users manage the increasing complexity of online tasks. The emergent reasoning ability in large language models offer a promising path for next-generation on-device AI agents. However, deploying full-scale Large Language Models (LLMs) on resource-limited local devices is challenging. In this paper, we propose Division-of-Thoughts (DoT), a collaborative reasoning framework leveraging the synergy between locally deployed Smaller-scale Language Models (SLMs) and cloud-based LLMs. DoT leverages a Task Decomposer to elicit the inherent planning abilities in language models to decompose user queries into smaller sub-tasks, which allows hybrid language models to fully exploit their respective strengths. Besides, DoT employs a Task Scheduler to analyze the pair-wise dependency of sub-tasks and create a dependency graph, facilitating parallel reasoning of sub-tasks and the identification of key steps. To allocate the appropriate model based on the difficulty of sub-tasks, DoT leverages a Plug-and-Play Adapter, which is an additional task head attached to the SLM that does not alter the SLM's parameters. To boost adapter's task allocation capability, we propose a self-reinforced training method that relies solely on task execution feedback. Extensive experiments on various benchmarks demonstrate that our DoT significantly reduces LLM costs while maintaining competitive reasoning accuracy. Specifically, DoT reduces the average reasoning time and API costs by 66.12% and 83.57%, while achieving comparable reasoning accuracy with the best baseline methods.