-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy patheval_instruction.py
75 lines (66 loc) · 5.95 KB
/
eval_instruction.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
# Instruction List For Causual Embedding Model
task2instruction = {
"FinancialPhraseBankClassification": "Classify the sentiment of a given finance text as either positive, negative, or neutral.",
"FinSentClassification": "Classify the sentiment of a given finance text as either positive, negative, or neutral.",
"FiQAClassification": "Perform aspect based financial sentiment classification.",
"SemEva2017Classification": "Classify the sentiment of a given finance text as either positive, negative, or neutral.",
"FLSClassification": "Classify the sentence into 'not-fls', 'specific fls', or 'non-specific fls' class.",
"ESGClassification": "Classify the following sentence into one of the 'environmental','social', 'governance', 'non-esg' classes.",
"FOMCClassification": "Classify the following sentence from FOMC into 'hawkish', 'dovish', or 'neutral' class.",
"FinancialFraudClassification": "Detecting financial fraud from the given text.",
"FinNSPClassification": "Perform financial sentiment classification.",
"FinChinaSentimentClassification": "Classify the sentiment of a given finance text as either positive, negative, or neutral.",
"FinFEClassification": "Classify the sentiment of a given financial social media text.",
"OpenFinDataSentimentClassification": "Classify the sentiment of a given finance text.",
"Weibo21Classification": "Classify fake news from the given text.",
"FiQA2018Retrieval": "Given a financial question, retrieve user replies that best answer the question.",
"FinanceBenchRetrieval": "Given a financial question, retrieve the related context.",
"HC3Retrieval": "Given a financial question, retrieve relevant passages that answer the query.",
"Apple10KRetrieval": "Given a financial question, retrieve the related context.",
"FinQARetrieval": "Given a financial question, retrieve user replies that best answer the question.",
"TATQARetrieval": "Given a financial question, retrieve user replies that best answer the question.",
"USNewsRetrieval": "Given a financial question, retrieve documents that can help answer the question.",
"TradeTheEventEncyclopediaRetrieval": "Given a financial term, retrieve the related context.",
"TradeTheEventNewsRetrieval": "Given a financial question, retrieve user replies that best answer the question.",
"TheGoldmanEnRetrieval": "Given a financial term, retrieve the related context.",
"FinTruthQARetrieval": "Given a financial question, retrieve user replies that best answer the question.",
"FinEvaRetrieval": "Given a financial question, retrieve user replies that best answer the question.",
"AlphaFinRetrieval": "Given a financial question, retrieve user replies that best answer the question.",
"DISCFinLLMRetrieval": "Given a financial question, retrieve documents that answer the query.",
"DISCFinLLMComputingRetrieval": "Given a financial question, retrieve the best answer.",
"DuEEFinRetrieval": "Given a financial question, retrieve documents that can help answer the question.",
"SmoothNLPRetrieval": "Given a financial question, retrieve documents that can help answer the question.",
"THUCNewsRetrieval": "Given a financial question, retrieve documents that can help answer the question.",
"FinEvaEncyclopediaRetrieval": "Given a financial term, retrieve the related context.",
"TheGoldmanZhRetrieval": "Given a financial term, retrieve the related context.",
"MInDS14EnClustering": "Identify the main category of the intention for the given text.",
"ComplaintsClustering": "Identify the main category of the consumer complaint.",
"PiiClustering": "Cluster the given text based on the personally identifiable information.",
"FinanceArxivS2SClustering": "Identify the main category of finance papers based on the titles",
"FinanceArxivP2PClustering": "Identify the main category of finance papers based on the abstracts.",
"WikiCompany2IndustryClustering": "Identify industries from company descriptions.",
"MInDS14ZhClustering": "Identify the main category of the intention for the given text.",
"FinNLClustering": "Identify the main category of the given finance news.",
"CCKS2022Clustering": "Identify the main event of the given text.",
"CCKS2020Clustering": "Identify the main event of the given text.",
"CCKS2019Clustering": "Identify the main event of the given text.",
"FinFactReranking": "Given a financial question, retrieve documents that answer the query.",
"FiQA2018Reranking": "Given a financial question, retrieve documents that can help answer the question.",
"HC3Reranking": "Given a financial question, retrieve relevant passages that answer the query.",
"FinEvaReranking": "Given a financial question, retrieve user replies that best answer the question.",
"DISCFinLLMReranking": "Given a financial query, retrieve the related context.",
"FinSTS": "Detecting Subtle Semantic Shifts in Financial Narratives.",
"FINAL": "Retrieve semantically similar finance text.",
"AFQMC": "Retrieve semantically similar finance text.",
"BQCorpus": "Retrieve semantically similar finance text.",
"Ectsum": "Given a news text, retrieve other semantically similar summaries",
"FINDsum": "Given a finance document, retrieve other semantically similar summaries",
"FNS2022sum": "Given a 10K document, retrieve other semantically similar summaries",
"FiNNAsum": "Given a news text, retrieve other semantically similar summaries",
"FinEvaHeadlinesum": "Given a finance document, retrieve other semantically similar summaries",
"FinEvasum": "Given a finance document, retrieve other semantically similar summaries",
"HeadlineACPairClassification": "Classify the sentiment of a given finance text.",
"HeadlinePDDPairClassification": "Classify the sentiment of a given finance text.",
"HeadlinePDUPairClassification": "Classify the sentiment of a given finance text.",
"AFQMCPairClassification": "Matching the semantically similar questions.",
}