Flora

15K posts

Flora banner
Flora

Flora

@flora_38657

Chinese-English translator

Katılım Temmuz 2013
112 Takip Edilen3.4K Takipçiler
Sabitlenmiş Tweet
Flora
Flora@flora_38657·
I am very excited to announce that our wiki now has its own domain‼️ Please update your bookmarks to zhang-zhehan.com We are still working on moving the info in our Twitters onto the wiki. New content is being added every day so check back often! #ZhangZhehan #张哲瀚
English
36
155
349
0
Flora retweetledi
老娘是你姑奶奶!
老娘是你姑奶奶!@QuelleVousNo2·
Happy birthday Gong Jun. This year I'm donating to the Red Cross in Hong Kong to aid the recovery of the Tai Po fire. If you're interested, you can also make a donation here. (Or donate to another charity of your choice.) donation.redcross.org.hk/onlinedonation…
English
4
14
74
9.4K
Flora
Flora@flora_38657·
Why are there so many editions of Harper's Bazaar anyway? It turns out that they license the use of the magazine name to local publishers. In other words, publication decisions have nothing to do with the parent company. hearstmagazinesinternational.com/licensing/
English
1
12
41
1.6K
Flora
Flora@flora_38657·
See anything familiar? The Harper's Bazaar link goes to Bazaar Vietnam, the same magazine that #ZhangSanjian was in.
Flora tweet mediaFlora tweet mediaFlora tweet media
English
1
14
41
1.7K
Flora
Flora@flora_38657·
Did you know that anybody can pay to be featured on fashion magazines? This PR agency, for $4000, can get you on 3 different publications: badenbower.com
English
6
21
64
5.8K
Flora retweetledi
Mirokany
Mirokany@mirokany·
I’ve completed the promised analysis comparing the speech patterns of Zhang Zhehan (ZZH) and Zhang Sanjian (ZSJ). I had far-reaching plans, but in the end I settled on standard tests, since they had already shown quite striking differences in their speech characteristics. But let's take things in order. For the analysis, I selected about 184 texts by ZZH (whatever I was able to find.) from both IG and Weibo and about 400 texts by ZSJ from IG. The difference in the corpora is large, but it does not play a special role, below I will explain why. At the first stage, I conducted a frequency analysis of words, and then identified words characteristic of the speech of ZZH and ZSJ. I got a rather impressive result – the two data sets simply had nothing in common with each other, except for common words such as "you", "we", "happiness", "day" and so on. Such words are found in the speech of literally every person. The next thing that caught my eye was the abundance of Anglicisms in ZSJ's speech and brand names in ZZH, such as Pandora, Coca-Cola, Maybelline. Here you need to understand that the model does not simply output the words that are used most often, but those that it considers a significant difference. ZZH is an actor who promoted certain brands, so it is not surprising that the model decided to highlight his professional vocabulary. On the other hand, ZSJ is clearly focused on an international audience, which explains the high level of English words in his corpus. I could have gotten such a counterargument, but since I was the first person to put it forward, for the sake of purity of the experiment I excluded obvious lexical markers from the analysis. The result has hardly changed. In P1 you can see the final version of the frequency analysis, as well as the count of unique words for each data set and the number of matches. What do these data tell us? 1. ZSJ has more unique words and seems to have a larger vocabulary, but this is not necessarily a good thing. This could be a marker of either AI text generation or that the texts are written by different people. 2. 253 words from the two sets match, which is very few. This already tells us that these are clearly two different people. 3. matching ZZH → ZSJ = 23.58% means that only 23.58% of ZZH words are used by ZSJ. 4. matching ZSJ → ZZH = 14.72% means that only 15% of ZSJ's vocabulary appears in ZZH texts. In both cases, the overlap between the two corpora is very low. If we were talking about one person, it would be logical to expect at least 40-60% overlap in vocabulary, taking into account all the aggravating circumstances, such as a change in activity, relocation, and psychological state. 5. If you look at the frequency analysis of words, ZZH's vocabulary is more reflective, mature, and specific, associated with struggle, challenges, traumatic or strong events. There is emotional depth in this. On the contrary, ZSJ's vocabulary is more superficial, saturated with streamlined emotional phrases, many repeating positive-motivating words. An influencer or SMM specialist could express themselves this way. Next, let's look at the frequency of parts of speech (POS). As I already said, the size of the ZZH and ZSJ corpora is not balanced in terms of the number of texts. I also said that this is not particularly important, and here's why. As you can see from the graph on P4, ZSJ has a lot of non-morphemic symbols, almost three times more than ZZH. With an imbalance in the data, a normal gap would be considered if the gap were only 2 times. What does this mean? This means that ZSJ often uses emoji and other symbols such as punctuation marks, which he uses as a logical conclusion to his thoughts. This is a fairly strong marker of artificiality of speech. Also, ZSJ's speech is almost 5 times more shows a much higher level of English words than ZZH's. That is, he literally dilutes Chinese speech with English, which was not typical of ZZH. Also, ZZH has the most frequent pattern of particle + noun + function word. Quite a common grammatical construction. By comparison, in ZSJ's texts, English words and non-morphemic symbols are used very often and in stable patterns. Thus, ZSJ's style is technically simpler, with frequent repeating patterns and high noise from emoji and English inserts. ZZH uses more varied and literary POS constructions, closer to the diary style. Conclusion: these two lexical corpora simply cannot belong to the same person. ZZH's speech patterns seem more stable, despite the fact that the texts are scattered over time and reflect a decade of his life. On the other hand, ZSJ seems more chaotic and unstable, which greatly strengthens my suspicion that the account on his behalf is run by different people, or is used by AI to generate texts. Finally, I would like to thank all the fans who took screenshots from Zhang Zhehan's Weibo and shared them online. Thank you guys, you are amazing and have helped a lot. This analysis is the result of your hard work. Without you, this research would not even be possible. 谢谢。
Mirokany tweet mediaMirokany tweet mediaMirokany tweet mediaMirokany tweet media
English
4
43
133
8.7K
Flora retweetledi
Mirokany
Mirokany@mirokany·
The IG post about Yale looked suspicious from the start. Those name tags with Chinese text were the first red flag. Then I noticed something else: no hashtags, no mention of Yale at all. Odd, right? Later I checked the official Yale SOM pages. Here’s what I found: 1.The name tags don’t match Yale’s style. Yale only prints names in English, the tags are usually laminated, and there’s no logo on them. 2.Yale’s Instagram shows nothing about this “event.” Only graduation posts at that time. 3.Yale’s calendar listed just 3 events on those dates: 2 were online, 1 in Miami. Nothing on campus. Classic scam move: drop a big name and hope no one checks the details.
Mirokany tweet mediaMirokany tweet media
English
4
14
51
4.5K
Flora
Flora@flora_38657·
Happy birthday #ZhangZhehan! 日月更替,山河依旧,我们永远都在♥️💙 张老师生日快乐啊!!!
中文
0
16
81
1.2K
Flora retweetledi
.www.
.www.@www_fanart·
[聯合產出]念春归 拾新绿 "月亮下的對白 單純的像小孩 你有好幾次問我 那是什麼 這就是愛" [音樂] ❤️🌙💙 媽咪Happy Birthday!!!🎂🎂🎂 不知不覺又一年啦~ 祝你身體健康、平平安安、一切順利、生活美滿快樂~ 爸比媽咪每天都要好好的🥰 繼續等待💕 P.S. 母親節快樂~!🌷😆 #俊哲 #JunZhe
.www. tweet media.www. tweet media
中文
3
117
501
10K
Flora
Flora@flora_38657·
Who would be willing to lose 5 figures to sue a vocal #ZhangZhehan and #GongJun fan, gaining her personal information, engaging her in an annoying legal battle and forcing her to post a public apology, with the possibility that this might cause her to leave the fandom? 🤔
English
1
1
25
2K
Flora
Flora@flora_38657·
Also, fansites typically must keep cameras on hand for short notice shoots, like running to the airport. After all, it's not like they have access to their idol's entire schedule. But only people who know exactly when in the future they need their cameras can rent them out.
English
3
2
23
1.6K
Flora
Flora@flora_38657·
Here's the most recent laurel factory that the ZSJ scam posted on their website: accoladecompetition.org The address is a Paper Source in a strip mall near UCSD. The phone number is for a design agency. The website is unregistered.
Flora tweet mediaFlora tweet media
English
3
16
44
4.3K
Flora
Flora@flora_38657·
These awards cater to indie film people who literally have no opportunity to showcase their projects & desperately need something on their CV. For someone with an established fanbase & industry experience, there is no purpose in collecting paid awards.
English
1
4
30
1.8K
Flora
Flora@flora_38657·
When the #Zhangsanjian scam first started going after film "awards", I signed up for an account on filmfreeway, the platform that you use to apply. And I learned there are literally countless of these awards/festivals. Very very few are actually competitive/reputable.
English
1
19
65
4.2K