Prompt engineering, as an efficient and effective way to leverage Large Language Models (LLM), has drawn a lot of attention from the research community. The existing research primarily emphasizes the importance of adapting prompts to specific tasks, rather than specific LLMs. However, a good prompt is not solely defined by its wording, but also binds to the nature of the LLM in question. In this work, we first quantitatively demonstrate that different prompts should be adapted to different LLMs to enhance their capabilities across various downstream tasks in NLP. Then we novelly propose a model-adaptive prompt optimizer (MAPO) method that optimizes the original prompts for each specific LLM in downstream tasks. Extensive experiments indicate that the proposed method can effectively refine prompts for an LLM, leading to significant improvements over various downstream tasks.

Parameter-efficient fine-tuning (PEFT) is crucial for customizing Large Language Models (LLMs) with constrained resources. Although there have been various PEFT methods for dense-architecture LLMs, PEFT for sparse-architecture LLMs is still underexplored. In this work, we study the PEFT method for LLMs with the Mixture-of-Experts (MoE) architecture and the contents of this work are mainly threefold: (1) We investigate the dispersion degree of the activated experts in customized tasks, and found that the routing distribution for a specific task tends to be highly concentrated, while the distribution of activated experts varies significantly across different tasks. (2) We propose Expert-Specialized Fine-Tuning, or ESFT, which tunes the experts most relevant to downstream tasks while freezing the other experts and modules; experimental results demonstrate that our method not only improves the tuning efficiency, but also matches or even surpasses the performance of full-parameter fine-tuning. (3) We further analyze the impact of the MoE architecture on expert-specialized fine-tuning. We find that MoE models with finer-grained experts are more advantageous in selecting the combination of experts that are most relevant to downstream tasks, thereby enhancing both the training efficiency and effectiveness. Our code is available at https://github.com/deepseek-ai/ESFT.

Large Language Models (LLMs) are powerful zero-shot assessors used in real-world situations such as assessing written exams and benchmarking systems. Despite these critical applications, no existing work has analyzed the vulnerability of judge-LLMs to adversarial manipulation. This work presents the first study on the adversarial robustness of assessment LLMs, where we demonstrate that short universal adversarial phrases can be concatenated to deceive judge LLMs to predict inflated scores. Since adversaries may not know or have access to the judge-LLMs, we propose a simple surrogate attack where a surrogate model is first attacked, and the learned attack phrase then transferred to unknown judge-LLMs. We propose a practical algorithm to determine the short universal attack phrases and demonstrate that when transferred to unseen models, scores can be drastically inflated such that irrespective of the assessed text, maximum scores are predicted. It is found that judge-LLMs are significantly more susceptible to these adversarial attacks when used for absolute scoring, as opposed to comparative assessment. Our findings raise concerns on the reliability of LLM-as-a-judge methods, and emphasize the importance of addressing vulnerabilities in LLM assessment methods before deployment in high-stakes real-world scenarios.

Large Language Models (LLMs) have significantly advanced natural language processing tasks such as machine translation, text generation, and sentiment analysis. However, their large size, often consisting of billions of parameters, poses challenges for storage, computation, and deployment, particularly in resource-constrained environments like mobile devices and edge computing platforms. Additionally, the key-value (k-v) cache used to speed up query processing requires substantial memory and storage, exacerbating these challenges. Vector databases have emerged as a crucial technology to efficiently manage and retrieve the high-dimensional vectors produced by LLMs, facilitating faster data access and reducing computational demands. Effective compression and quantization techniques are essential to address these challenges, as they reduce the memory footprint and computational requirements without significantly compromising performance. Traditional methods that uniformly map parameters to compressed spaces often fail to account for the uneven distribution of parameters, leading to considerable accuracy loss. Therefore, innovative approaches are needed to achieve better compression ratios while preserving model performance. In this work, we propose HERA, a novel algorithm that employs heuristic Element Replacement for compressing matrix. HERA systematically replaces elements within the model using heuristic methods, which simplifies the structure of the model and makes subsequent compression more effective. By hierarchically segmenting, compressing, and reorganizing the matrix dataset, our method can effectively reduce the quantization error to 12.3% of the original at the same compression ratio.

Large Language Models (LLMs) have sparked substantial interest and debate concerning their potential emergence of Theory of Mind (ToM) ability. Theory of mind evaluations currently focuses on testing models using machine-generated data or game settings prone to shortcuts and spurious correlations, which lacks evaluation of machine ToM ability in real-world human interaction scenarios. This poses a pressing demand to develop new real-world scenario benchmarks. We introduce NegotiationToM, a new benchmark designed to stress-test machine ToM in real-world negotiation surrounding covered multi-dimensional mental states (i.e., desires, beliefs, and intentions). Our benchmark builds upon the Belief-Desire-Intention (BDI) agent modeling theory and conducts the necessary empirical experiments to evaluate large language models. Our findings demonstrate that NegotiationToM is challenging for state-of-the-art LLMs, as they consistently perform significantly worse than humans, even when employing the chain-of-thought (CoT) method.

Controllability and proactivity are crucial properties of autonomous conversational agents (CAs). Controllability requires the CAs to follow the standard operating procedures (SOPs), such as verifying identity before activating credit cards. Proactivity requires the CAs to guide the conversation towards the goal during user uncooperation, such as persuasive dialogue. Existing research cannot be unified with controllability, proactivity, and low manual annotation. To bridge this gap, we propose a new framework for planning-based conversational agents (PCA) powered by large language models (LLMs), which only requires humans to define tasks and goals for the LLMs. Before conversation, LLM plans the core and necessary SOP for dialogue offline. During the conversation, LLM plans the best action path online referring to the SOP, and generates responses to achieve process controllability. Subsequently, we propose a semi-automatic dialogue data creation framework and curate a high-quality dialogue dataset (PCA-D). Meanwhile, we develop multiple variants and evaluation metrics for PCA, e.g., planning with Monte Carlo Tree Search (PCA-M), which searches for the optimal dialogue action while satisfying SOP constraints and achieving the proactive of the dialogue. Experiment results show that LLMs finetuned on PCA-D can significantly improve the performance and generalize to unseen domains. PCA-M outperforms other CoT and ToT baselines in terms of conversation controllability, proactivity, task success rate, and overall logical coherence, and is applicable in industry dialogue scenarios. The dataset and codes are available at XXXX.

Despite the rapid development of large language models (LLMs) for the Korean language, there remains an obvious lack of benchmark datasets that test the requisite Korean cultural and linguistic knowledge. Because many existing Korean benchmark datasets are derived from the English counterparts through translation, they often overlook the different cultural contexts. For the few benchmark datasets that are sourced from Korean data capturing cultural knowledge, only narrow tasks such as bias and hate speech detection are offered. To address this gap, we introduce a benchmark of Cultural and Linguistic Intelligence in Korean (CLIcK), a dataset comprising 1,995 QA pairs. CLIcK sources its data from official Korean exams and textbooks, partitioning the questions into eleven categories under the two main categories of language and culture. For each instance in CLIcK, we provide fine-grained annotation of which cultural and linguistic knowledge is required to answer the question correctly. Using CLIcK, we test 13 language models to assess their performance. Our evaluation uncovers insights into their performances across the categories, as well as the diverse factors affecting their comprehension. CLIcK offers the first large-scale comprehensive Korean-centric analysis of LLMs' proficiency in Korean culture and language.

MetaDesigner revolutionizes artistic typography synthesis by leveraging the strengths of Large Language Models (LLMs) to drive a design paradigm centered around user engagement. At the core of this framework lies a multi-agent system comprising the Pipeline, Glyph, and Texture agents, which collectively enable the creation of customized WordArt, ranging from semantic enhancements to the imposition of complex textures. MetaDesigner incorporates a comprehensive feedback mechanism that harnesses insights from multimodal models and user evaluations to refine and enhance the design process iteratively. Through this feedback loop, the system adeptly tunes hyperparameters to align with user-defined stylistic and thematic preferences, generating WordArt that not only meets but exceeds user expectations of visual appeal and contextual relevance. Empirical validations highlight MetaDesigner's capability to effectively serve diverse WordArt applications, consistently producing aesthetically appealing and context-sensitive results.

The open-ended nature of language generation makes the evaluation of autoregressive large language models (LLMs) challenging. One common evaluation approach uses multiple-choice questions (MCQ) to limit the response space. The model is then evaluated by ranking the candidate answers by the log probability of the first token prediction. However, first-tokens may not consistently reflect the final response output, due to model's diverse response styles such as starting with "Sure" or refusing to answer. Consequently, MCQ evaluation is not indicative of model behaviour when interacting with users. But by how much? We evaluate how aligned first-token evaluation is with the text output along several dimensions, namely final option choice, refusal rate, choice distribution and robustness under prompt perturbation. Our results show that the two approaches are severely misaligned on all dimensions, reaching mismatch rates over 60%. Models heavily fine-tuned on conversational or safety data are especially impacted. Crucially, models remain misaligned even when we increasingly constrain prompts, i.e., force them to start with an option letter or example template. Our findings i) underscore the importance of inspecting the text output as well and ii) caution against relying solely on first-token evaluation.

This paper presents a novel approach to aligning large language models (LLMs) with individual human preferences, sometimes referred to as Reinforcement Learning from \textit{Personalized} Human Feedback (RLPHF). Given stated preferences along multiple dimensions, such as helpfulness, conciseness, or humor, the goal is to create an LLM without re-training that best adheres to this specification. Starting from specialized expert LLMs, each trained for one such particular preference dimension, we propose a black-box method that merges their outputs on a per-token level. We train a lightweight Preference Control Model (PCM) that dynamically translates the preference description and current context into next-token prediction weights. By combining the expert models' outputs at the token level, our approach dynamically generates text that optimizes the given preference. Empirical tests show that our method matches or surpasses existing preference merging techniques, providing a scalable, efficient alternative to fine-tuning LLMs for individual personalization.