2023
DOI: 10.2139/ssrn.4372889
|View full text |Cite
Preprint
|
Sign up to set email alerts
|

Chatgpt: Jack of All Trades, Master of None

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
2
1

Citation Types

1
66
0
1

Year Published

2023
2023
2024
2024

Publication Types

Select...
5
2
1
1

Relationship

0
9

Authors

Journals

citations
Cited by 70 publications
(68 citation statements)
references
References 54 publications
1
66
0
1
Order By: Relevance
“…Pre-trained large language models (LLMs; e.g., GPT-3.5, ChatGPT and GPT-4), which are performed through chatting (or asking) with it, have obtained promising results on various natural language understanding (NLU) and natural language generation (NLG) downstream tasks (Ouyang et al, 2022;Kocoń et al, 2023;Qin et al, 2023;Rao et al, 2023;Bang et al, 2023;Zuccon and Koopman, 2023). For example, Zhong et al (2023) show that Chat-GPT can attain the comparable understanding ability to some fine-tuned BERT-style models on NLU arXiv:2303.04048v2 [cs.CL] 25 Apr 2023 tasks while failing to surpass current task-specific NLU models.…”
Section: Introductionmentioning
confidence: 99%
“…Pre-trained large language models (LLMs; e.g., GPT-3.5, ChatGPT and GPT-4), which are performed through chatting (or asking) with it, have obtained promising results on various natural language understanding (NLU) and natural language generation (NLG) downstream tasks (Ouyang et al, 2022;Kocoń et al, 2023;Qin et al, 2023;Rao et al, 2023;Bang et al, 2023;Zuccon and Koopman, 2023). For example, Zhong et al (2023) show that Chat-GPT can attain the comparable understanding ability to some fine-tuned BERT-style models on NLU arXiv:2303.04048v2 [cs.CL] 25 Apr 2023 tasks while failing to surpass current task-specific NLU models.…”
Section: Introductionmentioning
confidence: 99%
“…The paper also does not provide a detailed analysis of the factors affecting the translation performance of ChatGPT. Kocon et al [43] evaluated the capabilities of the ChatGPT on 25 diverse analytical NLP tasks, most of which are subjective in nature, including sentiment analysis, emotion recognition, offensiveness and stance detection, natural language inference, word sense disambiguation, linguistic acceptability, and question answering. The authors automated ChatGPT's querying process and analyzed more than 38k responses, comparing its results with state-of-the-art solutions.…”
Section: A Natural Language Processingmentioning
confidence: 99%
“…This interactive nature allows users to leverage the model for a wide range of NLP tasks using conversational texts, which is called a prompt. The methodology of crafting prompts to obtain desired outputs from the off-the-shelf model without additional dataset (i.e, zero-shot setting) is referred to as prompt engineering [10]. Additionally, there is a methodology that employs a dataset of expected input-output pairs to fine-tune the model's behavior.…”
Section: A Chatgpt and Promptsmentioning
confidence: 99%
“…In recent years, the advancements in Large Language Model (LLM), notably models like ChatGPT, have shown promise in various Natural Language Processing (NLP) tasks, including emotion recognition. Early studies have demonstrated the capability of ChatGPT in basic sentiment analysis tasks where the primary goal is to distinguish between positive and negative sentiments [7]- [10]. Beyond this binary classification a few reports delved into a nuanced understanding of emotions like joy, sadness, anger, and surprise and showed that ChatGPT shows reasonable performance in such detailed emotion analysis [5], [11] in a zero-shot and few-shot prompting conditions.…”
Section: Introductionmentioning
confidence: 99%