Spaces:
Running
Running
| from dataclasses import dataclass | |
| from enum import Enum | |
| NUM_FEWSHOT = 0 # Change with your few shot | |
| # --------------------------------------------------- | |
| TITLE = """<h1>🇹🇭 Thai LLM Leaderboard</h1>""" | |
| # <a href="url"></a> | |
| INTRODUCTION_TEXT = """ | |
| The Thai LLM Leaderboard 🇹🇭 aims to standardize evaluation methods for large language models (LLMs) in the Thai language, building on <a href="https://github.com/SEACrowd">SEACrowd</a>. | |
| As part of an open community project, we welcome you to submit new evaluation tasks or models. | |
| This leaderboard is developed in collaboration with <a href="https://www.scb10x.com">SCB 10X</a>, <a href="https://www.vistec.ac.th/">VISTEC</a>, and <a href="https://github.com/SEACrowd">SEACrowd</a>. | |
| """ | |
| LLM_BENCHMARKS_TEXT = f""" | |
| The leaderboard currently consists of the following benchmarks: | |
| - <b>Exam</b> | |
| - <a href="https://huggingface.co/datasets/scb10x/thai_exam">ThaiExam</a>: ThaiExam is a Thai language benchmark based on examinations for high-school students and investment professionals in Thailand. | |
| - <a href="https://arxiv.org/abs/2306.05179">M3Exam</a>: M3Exam is a novel benchmark sourced from authentic and official human exam questions for evaluating LLMs in a multilingual, multimodal, and multilevel context. This leaderboard uses the Thai subset of M3Exam. | |
| - <b>LLM-as-a-Judge</b> | |
| - <a href="https://huggingface.co/datasets/ThaiLLM-Leaderboard/mt-bench-thai">Thai MT-Bench</a>: A Thai version of <a href="https://arxiv.org/abs/2306.05685">MT-Bench</a> developed specially by VISTEC for probing Thai generative skills using the LLM-as-a-judge method. | |
| - <b>NLU</b> | |
| - <a href="https://huggingface.co/datasets/facebook/belebele">Belebele</a>: Belebele is a multiple-choice machine reading comprehension (MRC) dataset spanning 122 language variants, where the Thai subset is used in this leaderboard. | |
| - <a href="https://huggingface.co/datasets/facebook/xnli">XNLI</a>: XNLI is an evaluation corpus for language transfer and cross-lingual sentence classification in 15 languages. This leaderboard uses the Thai subset of this corpus. | |
| - <a href="https://huggingface.co/datasets/cambridgeltl/xcopa">XCOPA</a>: XCOPA is a corpus of translated and re-annotated English COPA, covers 11 languages. This is designed to measure the commonsense reasoning ability in non-English languages. This leaderboard uses the Thai subset of this corpus. | |
| - <a href="https://huggingface.co/datasets/pythainlp/wisesight_sentiment">Wisesight</a>: Wisesight sentiment analysis corpus contains social media messages in the Thai language with sentiment labels. | |
| - <b>NLG</b> | |
| - <a href="https://huggingface.co/datasets/csebuetnlp/xlsum">XLSum</a>: XLSum is a comprehensive and diverse dataset comprising 1.35 million professionally annotated article-summary pairs from the BBC. This corpus evaluates the summarization performance in non-English languages, and this leaderboard uses the Thai subset. | |
| - <a href="https://huggingface.co/datasets/SEACrowd/flores200">Flores200</a>: FLORES is a machine translation benchmark dataset used to evaluate translation quality between English and low-resource languages. This leaderboard uses the Thai subset of Flores200. | |
| - <a href="https://huggingface.co/datasets/iapp/iapp_wiki_qa_squad">iapp Wiki QA Squad</a>: iapp Wiki QA Squad is an extractive question-answering dataset derived from Thai Wikipedia articles. | |
| <b>Metric Implementation Details</b>: | |
| - BLEU is calculated using flores200's tokenizer using HuggingFace `evaluate` <a href="https://huggingface.co/spaces/evaluate-metric/sacrebleu">implementation</a>. | |
| - ROUGEL is calculated using PyThaiNLP newmm tokenizer and HuggingFace `evaluate` <a href="https://huggingface.co/spaces/evaluate-metric/rouge">implementation</a>. | |
| - LLM-as-a-judge rating is based on OpenAI's gpt-4o-2024-05-13 using the prompt defined in <a href="https://github.com/lm-sys/FastChat/blob/main/fastchat/llm_judge/data/judge_prompts.jsonl">lmsys MT-Bench</a>. | |
| <b>Reproducibility</b>: | |
| - For the reproducibility of results, we have open-sourced the evaluation pipeline. Please check out the repository <a href="https://github.com/scb-10x/seacrowd-eval">seacrowd-experiments</a>. | |
| <b>Acknowledgements</b>: | |
| - We are grateful to previous open-source projects that released datasets, tools, and knowledge. We thank community members for tasks and model submissions. To contribute, please see the submit tab. | |
| """ | |
| CITATION_BUTTON_LABEL = "Copy the following snippet to cite these results" | |
| CITATION_BUTTON_TEXT = r"""@misc{thaillm-leaderboard, | |
| author={SCB 10X and VISTEC and SEACrowd}, | |
| title={Thai LLM Leaderboard}, | |
| year={2024}, | |
| publisher={Hugging Face}, | |
| url={https://huggingface.co/spaces/ThaiLLM-Leaderboard/leaderboard} | |
| }""" |