Prince Kumar

Mehant Kammakomati, Sameer Pimparkhede, Srikanth Tamilselvam et al.

Recent work shows Large Language Models (LLMs) struggle to understand natural language constraints for various text generation tasks in zero- and few-shot settings. While, in the code domain, there is wide usage of constraints in code format to maintain the integrity of code written in Domain-Specific Languages (DSLs) like JSON and YAML which are widely used for system-level programming tasks in enterprises. Given that LLMs are increasingly used for system-level code tasks, evaluating if they can comprehend these code constraints is crucial. However, no work has been done to evaluate their controllability over code constraints. Hence, we introduce ConCodeEval, a first-of-its-kind benchmark having two novel tasks for code constraints across five representations. Our findings suggest that language models struggle with code constraints. Code languages that perform excellently for normal code tasks do not perform well when the same languages represent fine-grained constraints.

Sameer Pimparkhede, Mehant Kammakomati, Srikanth Tamilselvam et al.

Recent developments show that Large Language Models (LLMs) produce state-of-the-art performance on natural language (NL) to code generation for resource-rich general-purpose languages like C++, Java, and Python. However, their practical usage for structured domain-specific languages (DSLs) such as YAML, JSON is limited due to domain-specific schema, grammar, and customizations generally unseen by LLMs during pre-training. Efforts have been made to mitigate this challenge via in-context learning through relevant examples or by fine-tuning. However, it suffers from problems, such as limited DSL samples and prompt sensitivity but enterprises maintain good documentation of the DSLs. Therefore, we propose DocCGen, a framework that can leverage such rich knowledge by breaking the NL-to-Code generation task for structured code languages into a two-step process. First, it detects the correct libraries using the library documentation that best matches the NL query. Then, it utilizes schema rules extracted from the documentation of these libraries to constrain the decoding. We evaluate our framework for two complex structured languages, Ansible YAML and Bash command, consisting of two settings: Out-of-domain (OOD) and In-domain (ID). Our extensive experiments show that DocCGen consistently improves different-sized language models across all six evaluation metrics, reducing syntactic and semantic errors in structured code. We plan to open-source the datasets and code to motivate research in constrained code generation.

Rudra Murthy, Praveen Venkateswaran, Prince Kumar et al.

LLM evaluation benchmarks have traditionally separated the testing of knowledge/reasoning capabilities from instruction following. In this work, we study the interaction between knowledge and instruction following, and observe that LLMs struggle to follow simple answer modifying instructions, and are also distracted by instructions that should have no bearing on the original knowledge task answer. We leverage existing multiple-choice answer based knowledge benchmarks and apply a set of simple instructions which include manipulating text (eg.: change case), numeric quantities (eg.: increase value, change formatting), operate on lists (eg.: sort answer candidates) and distractor instructions (eg.: change case of numeric answers). We evaluate models at varying parameter sizes (1B-405B) from different model families and find that, surprisingly, all models report a significant drop in performance on such simple task compositions. While large-sized and frontier models report performance drops of 40-50%, in small and medium sized models the drop is severe (sometimes exceeding 80%). Our results highlight a limitation in the traditional separation of knowledge/reasoning and instruction following, and suggest that joint-study of these capabilities are important. We release our benchmark dataset, evaluation framework code, and results for future work.

Varad Bhatnagar, Prince Kumar, Sairam Moghili et al.

Recently the NLP community has started showing interest towards the challenging task of Hostile Post Detection. This paper present our system for Shared Task at Constraint2021 on "Hostile Post Detection in Hindi". The data for this shared task is provided in Hindi Devanagari script which was collected from Twitter and Facebook. It is a multi-label multi-class classification problem where each data instance is annotated into one or more of the five classes: fake, hate, offensive, defamation, and non-hostile. We propose a two level architecture which is made up of BERT based classifiers and statistical classifiers to solve this problem. Our team 'Albatross', scored 0.9709 Coarse grained hostility F1 score measure on Hostile Post Detection in Hindi subtask and secured 2nd rank out of 45 teams for the task. Our submission is ranked 2nd and 3rd out of a total of 156 submissions with Coarse grained hostility F1 score of 0.9709 and 0.9703 respectively. Our fine grained scores are also very encouraging and can be improved with further finetuning. The code is publicly available.