Zobrazeno 1 - 10
of 20
pro vyhledávání: '"Wang, Sida"'
Autor:
Jain, Naman, Han, King, Gu, Alex, Li, Wen-Ding, Yan, Fanjia, Zhang, Tianjun, Wang, Sida, Solar-Lezama, Armando, Sen, Koushik, Stoica, Ion
Large Language Models (LLMs) applied to code-related applications have emerged as a prominent field, attracting significant interest from both academia and industry. However, as new and improved LLMs are developed, existing evaluation benchmarks (e.g
Externí odkaz:
http://arxiv.org/abs/2403.07974
We introduce Syntax-Aware Fill-In-the-Middle (SAFIM), a new benchmark for evaluating Large Language Models (LLMs) on the code Fill-in-the-Middle (FIM) task. This benchmark focuses on syntax-aware completions of program structures such as code blocks
Externí odkaz:
http://arxiv.org/abs/2403.04814
Autor:
Gu, Alex, Rozière, Baptiste, Leather, Hugh, Solar-Lezama, Armando, Synnaeve, Gabriel, Wang, Sida I.
We present CRUXEval (Code Reasoning, Understanding, and eXecution Evaluation), a benchmark consisting of 800 Python functions (3-13 lines). Each function comes with an input-output pair, leading to two natural tasks: input prediction and output predi
Externí odkaz:
http://arxiv.org/abs/2401.03065
Autor:
Wang, Sida I.
The striking ability of unsupervised word translation has been demonstrated with the help of word vectors / pretraining; however, they require large amounts of data and usually fails if the data come from different domains. We propose coocmap, a meth
Externí odkaz:
http://arxiv.org/abs/2305.14200
Autor:
Ni, Ansong, Iyer, Srini, Radev, Dragomir, Stoyanov, Ves, Yih, Wen-tau, Wang, Sida I., Lin, Xi Victoria
The advent of large language models trained on code (code LLMs) has led to significant progress in language-to-code generation. State-of-the-art approaches in this area combine LLM decoding with sample pruning and reranking using test cases or heuris
Externí odkaz:
http://arxiv.org/abs/2302.08468
Autor:
Zhang, Tianyi, Yu, Tao, Hashimoto, Tatsunori B., Lewis, Mike, Yih, Wen-tau, Fried, Daniel, Wang, Sida I.
Sampling diverse programs from a code language model and reranking with model likelihood is a popular method for code generation but it is prone to preferring degenerate solutions. Inspired by collaborative programming, we propose Coder-Reviewer rera
Externí odkaz:
http://arxiv.org/abs/2211.16490
Autor:
Lin, Bill Yuchen, Wang, Sida, Lin, Xi Victoria, Jia, Robin, Xiao, Lin, Ren, Xiang, Yih, Wen-tau
Real-world natural language processing (NLP) models need to be continually updated to fix the prediction errors in out-of-distribution (OOD) data streams while overcoming catastrophic forgetting. However, existing continual learning (CL) problem setu
Externí odkaz:
http://arxiv.org/abs/2205.02014
Autor:
Fried, Daniel, Aghajanyan, Armen, Lin, Jessy, Wang, Sida, Wallace, Eric, Shi, Freda, Zhong, Ruiqi, Yih, Wen-tau, Zettlemoyer, Luke, Lewis, Mike
Code is seldom written in a single left-to-right pass and is instead repeatedly edited and refined. We introduce InCoder, a unified generative model that can perform program synthesis (via left-to-right generation) as well as editing (via infilling).
Externí odkaz:
http://arxiv.org/abs/2204.05999
Existing work in language grounding typically study single environments. How do we build unified models that apply across multiple environments? We propose the multi-environment Symbolic Interactive Language Grounding benchmark (SILG), which unifies
Externí odkaz:
http://arxiv.org/abs/2110.10661
Active learning (AL) algorithms may achieve better performance with fewer data because the model guides the data selection process. While many algorithms have been proposed, there is little study on what the optimal AL algorithm looks like, which wou
Externí odkaz:
http://arxiv.org/abs/2101.00977