Large language models (LLMs) have transformed code generation.However, most existing approaches focus on mainstream languages such as Python and Java, neglecting the Solidity language, the predominant programming language for Ethereum smart contracts.Due to the lack of adequate benchmarks for Solidity, LLMs’ ability to generate secure, cost-effective smart contracts remains unexplored.To fill this gap, we construct SolEval, the first repository-level benchmark designed for Solidity smart contract generation, to evaluate the performance of LLMs on Solidity.SolEval consists of 1,507 samples from 28 different repositories, covering 6 popular domains, providing LLMs with a comprehensive evaluation benchmark.Unlike the existing Solidity benchmark, SolEval not only includes complex function calls but also reflects the real-world complexity of the Ethereum ecosystem by incorporating Gas@k and Vul@k.We evaluate 16 LLMs on SolEval, and our results show that the best-performing LLM achieves only 26.29% Pass@10, highlighting substantial room for improvement in Solidity code generation by LLMs.Additionally, we conduct supervised fine-tuning (SFT) on Qwen-7B using SolEval, resulting in a significant performance improvement, with Pass@5 increasing from 16.67% to 58.33%, demonstrating the effectiveness of fine-tuning LLMs on our benchmark.We release our data and code at https://github.com/pzy2000/SolEval.

Domain knowledge is important to understand both the lexical and relational associations of words in natural language text, especially for domain-specific tasks like Natural Language Inference (NLI) in the medical domain, where due to the lack of a large annotated dataset such knowledge cannot be implicitly learned during training. However, because of the linguistic idiosyncrasies of clinical texts (e.g., shorthand jargon), solely relying on domain knowledge from an external knowledge base (e.g., UMLS) can lead to wrong inference predictions as it disregards contextual information and, hence, does not return the most relevant mapping. To remedy this, we devise a knowledge adaptive approach for medical NLI that encodes the premise/hypothesis texts by leveraging supplementary external knowledge, alongside the UMLS, based on the word contexts. By incorporating refined domain knowledge at both the lexical and relational levels through a multi-source attention mechanism, it is able to align the token-level interactions between the premise and hypothesis more effectively. Comprehensive experiments and case study on the recently released MedNLI dataset are conducted to validate the effectiveness of the proposed approach.

Variational encoder-decoders have achieved well-recognized performance in the dialogue generation task. Existing works simply assume the Gaussian priors of the latent variable, which are incapable of representing complex latent variables effectively. To address the issues, we propose to use the Dirichlet distribution with flexible structures to characterize the latent variables in place of the traditional Gaussian distribution, called Dirichlet Latent Variable Hierarchical Recurrent Encoder-Decoder model (Dir-VHRED). Based on which, we further find that there is redundancy among the dimensions of latent variable, and the lengths and sentence patterns of the responses can be strongly correlated to each dimension of the latent variable. Therefore, controllable responses can be generated through specifying the value of each dimension of the latent variable. Experimental results on benchmarks show that our proposed Dir-VHRED yields substantial improvements on negative log-likelihood, word-embedding-based and human evaluations.

Large enterprises, such as IBM, accumulate petabytes of free-text data within their organizations. To mine this big data, a critical ability is to enable meaningful question answering beyond keywords search. In this paper, we present a study on the characteristics and classification of IBM sales questions. The characteristics are analyzed both semantically and syntactically, from where a question classification guideline evolves. We adopted an enterprise level expert sourcing approach to gather questions, annotate questions based on the guideline and manage the quality of annotations via enhanced inter-annotator agreement analysis. We developed a question feature extraction system and experimented with rule-based, statistical and hybrid question classifiers. We share our annotated corpus of questions and report our experimental results. Statistical classifiers separately based on n-grams and hand-crafted rule features give reasonable macro-f1 scores at 61.7% and 63.1% respectively. Rule based classifier gives a macro-f1 at 77.1%. The hybrid classifier with n-gram and rule features using a second guess model further improves the macro-f1 to 83.9%.