28487
#ielts #toefl #gre #english_vocabulary #english
Language Log
Invisible text via Unicode tag characters
If you open this file in your browser, you'll see only an an left square bracket followed by a right square bracket, with nothing in between: http://languagelog.ldc.upenn.edu/myl/LanguageLogInTags0.png But if I run the file through a perl script that I wrote long ago to print out character codes and their names, I get
|[| 0x005B "LEFT SQUARE BRACKET"
|| 0xE004C "TAG LATIN CAPITAL LETTER L"
|| 0xE0061 "TAG LATIN SMALL LETTER A"
|| 0xE006E "TAG LATIN SMALL LETTER N"
|| 0xE0067 "TAG LATIN SMALL LETTER G"
|| 0xE0075 "TAG LATIN SMALL LETTER U"
|| 0xE0061 "TAG LATIN SMALL LETTER A"
|| 0xE0067 "TAG LATIN SMALL LETTER G"
|| 0xE0065 "TAG LATIN SMALL LETTER E"
|| 0xE0020 "TAG SPACE"
|| 0xE004C "TAG LATIN CAPITAL LETTER L"
|| 0xE006F "TAG LATIN SMALL LETTER O"
|| 0xE0067 "TAG LATIN SMALL LETTER G"
|]| 0x005D "RIGHT SQUARE BRACKET"
Or you can cut and paste the bracketed sequence into the ASCII Smuggler page, and you'll see this: http://languagelog.ldc.upenn.edu/myl/LanguageLogInTags1.png For an explanation, see Johann Rehberger's blog post "ASCII Smuggler Tool: Crafting Invisible Text and Decoding Hidden Codes", Embrace The Red 1/14/2024, which starts:
A few days ago Riley Goodside posted about an interesting discovery on how an LLM prompt injection can happen via invisible instructions in pasted text. This works by using a special set of Unicode code points from the Tags Unicode Block.
The proof-of-concept showed how a simple text contained invisible instructions that caused ChatGPT to invoke DALL-E to create an image.
The meaning of these “Tags” seems to have gone through quite some churn, from language tags to eventually being repurposed for some emojis. […]
The Tags Unicode Block mirrors ASCII and because it is often not rendered in the UI, the special text remains unnoticable to users… but LLMs interpret such text.
It appears that training data contained such characters and now tokenizers can deal with them!
For more recent coverage, see Dan Goodin, "Invisible text that AI chatbots understand and humans can’t? Yep, it’s a thing.", Ars Technica 10/14/2024, which explains Rehberger's "Copirate" hack:
What if there was a way to sneak malicious instructions into Claude, Copilot, or other top-name AI chatbots and get confidential data out of them by using characters large language models can recognize and their human users can’t? As it turns out, there was—and in some cases still is.
The invisible characters, the result of a quirk in the Unicode text encoding standard, create an ideal covert channel that can make it easier for attackers to conceal malicious payloads fed into an LLM. The hidden text can similarly obfuscate the exfiltration of passwords, financial information, or other secrets out of the same AI-powered bots. Because the hidden text can be combined with normal text, users can unwittingly paste it into prompts. The secret content can also be appended to visible text in chatbot output. […]
“The fact that GPT 4.0 and Claude Opus were able to really understand those invisible tags was really mind-blowing to me and made the whole AI security space much more interesting,” Joseph Thacker, an independent researcher and AI engineer at Appomni, said in an interview. “The idea that they can be completely invisible in all browsers but still readable by large language models makes [attacks] much more feasible in just about every area.”
Microsoft has mitigated the threat to Copilot, according to Ravie Lakshmanan, "Microsoft Fixes ASCII Smuggling Flaw That Enabled Data Theft from Microsoft 365 Copilot", The Hacker News 8/27/2024. I'm not sure where things stand with ChatGPT, Claude, Gemini, and so on. As far as I can tell, WordPress removes tag characters from the text that it stores and presents.
For more details about the Copilot exploit, see this video:
➖ @EngSkills ➖
Word of the Day
Word of the Day: concurrent
This word has appeared in 83 articles on NYTimes.com in the past year. Can you use it in a sentence?
➖ @EngSkills ➖
Phrasal Verb of the Day | Vocabulary | EnglishClub
stand out
If somebody stands out, they are easy to see because there is something unusual about the way they look or the way they behave.
➖ @EngSkills ➖
Word of the Day
cagey
Definition: (adjective) Showing self-interest and shrewdness in dealing with others.
Synonyms: canny, clever.
Usage: The jurors saw right through the cagey lawyer's attempt to divert their attention away from the evidence.
Discuss
➖ @EngSkills ➖
Funny Or Die (Youtube)
"Don't work with kids, animals, or drunks": How Drunk History Was Made
Yes, they were really drunk and it was real history.
Watch the full episode with Derek Waters here: https://youtu.be/t4TXV4PC2PA
Get all 10 episodes of season 1 now, and stay in touch for new episodes, news, and show extras: https://norby.link/ctdAJD
Subscribe now: https://www.youtube.com/c/funnyordie?sub_confirmation=1
Get more Funny Or Die
-------------------------------
Facebook: https://www.facebook.com/funnyordie
Twitter: https://twitter.com/funnyordie
Instagram: http://instagram.com/funnyordie
TikTok: funnyordie" rel="nofollow">https://www.tiktok.com/@funnyordie
➖ @EngSkills ➖
Language Log
Shaikh Zubayr
Sean Swanick, "Shaikh Zubayr", Duke University Libraries Blog, 4/13/2016:
A man lost at sea, having drifted far away from his native Iraqi lands, comes a shore in England. In due time he will be nicknamed the Bard of Avon but upon landing on the Saxon coast, his passport reportedly read: Shaikh Zubayr. A knowledgeable man with great writing prowess from a small town called Zubayr in Iraq. He came to be known in the West as Shakespeare and was given the first name of William. William Shakespeare of Zubayr.
Or at least this is loosely how a story goes about The Bard’s origins. It was purportedly first suggested by the famous Lebanese intellectual, Aḥmad Fāris al-Shidyāq and later popularised by the Iraqi intellectual Dr. Ṣafāʾ Khulūṣī. Khulūṣī in 1960 published an article entitled “The Study of Shakespeare” in al-Ma’rifa (1960) where he paid homage to the Bard while also expanding upon al-Shidyāq’s theory. This perplexing theory has generated much rebuttal and discussion. This theory rested upon “that most of Shakespeare’s language could be traced back to Classical Arabic…[t]o give one example : the Arabic adjective nabīl which means ‘noble’ and which occurs, naturally enough, throughout the plays and poems.” (“Shadow Language” in Ormsby, Eric L. 2011. Fine incisions: essays on poetry and place. Erin, Ont: Porcupine’s Quill.) The former Libyan dictator, Mu’ammar al-Qadhāfī is also reported to have supported the theory of Shaikh Zubayr.
See also Abdul Sattar Jawad, "Shakespeare in Baghdad", The Chronicle 12/2/2011, who spells it Sheikh Zbair.
I recall reading in one of Sir Richard Burton's works about his discussion with a Somali host, who argued that the English might be good at making machines, but they lacked the poetry for which the Somalis were notable. Burton countered by quoting Shakespeare, to which his host responded "And this Sheik Subeer, how many camels does he have?"
Or maybe I dreamed it, as I haven't been able to locate the passage. I thought it was in First Footsteps in East Africa, but it's not.
➖ @EngSkills ➖
Slang of the Day | Vocabulary | EnglishClub
pimp (1)
a manager of prostitutes
➖ @EngSkills ➖
Idiom of the Day
go Galt
To reduce, restrict, or cease one's work or productivity as a means of social protest against increased marginal tax rates, limits on tax deductions, or the use of tax income for purposes one finds morally objectionable. Taken from the name John Galt, a character in Ayn Rand's novel Atlas Shrugged. Primarily heard in US, South Africa. Watch the video
➖ @EngSkills ➖
Are you free now? I'll invite you to coffee.
I don't drink coffee, I only drink tea.
I have Chinese green tea, and I also have English black tea. Which would you like?
Green tea, black tea; I'll drink either.
Students cannot recognise any characters in this lesson when asked about them the next week. For homework, the students write characters on a word doc and send it to me so I feel like they would input using pinyin.
5/ Students are mostly young office workers so their goals are for job prospects and travel in China. In recent years, a lot of Chinese companies have entered Southeast Asia (e.g., Tiktok and Temu) and the management are mostly Chinese instead of locals. So office workers are motivated to learn Chinese. But because of the time constraint, they prioritise speaking and listening as it sees a quicker return (probably because they do not have to spend much effort in learning pinyin). Since grades do not matter, I cannot require much from them, and it mostly depends on how much effort (and time) they're willing to expend.
Next year I am planning to try teaching Chinese in offline centers (after I have graduated from Tsinghua) to see if it is a better alternative. But besides the "online restrictions", I feel like the student segment is also a big factor because office workers have too much on their plate (Also looking forward to approaching other student segments to test this hypothesis).
As I have repeatedly pointed out, students in Singapore (even those who are ethnic Chinese) are permitted to"write" their characters with computers and other digital devices. There's a world of difference between writing hanzi by hand and using electronic tools. The former is much, much harder — excruciatingly more difficult — than the latter.
If someone told you that you could become fully fluent in Mandarin without having to endure the agony of memorizing a single sinograph, much less to expend the months and years of toil required to master the Chinese writing system, would you do it?
P.S.: For typical human beings, there's no "quick tip" for learning to write Chinese characters. It's brute memorization the whole way. You have to spend time, lots and lots of it.
Selected readings
* "'They're not learning how to write characters!'" (11/5/21) — with a long bibliography learning how to read and write hanzi
* "Fluent bilingualism in Singapore" (5/28/19)
* "Characterless Sinitic" (9/1/21) — directly pertinent to this post; with a lengthy, helpful bibliography
➖ @EngSkills ➖
Slang of the Day | Vocabulary | EnglishClub
egosurf
to search for one's own name on the Internet
➖ @EngSkills ➖
Idiom of the Day
glimmer of hope
A minute indication that something may improve, succeed, or turn out for the best in the end. Watch the video
➖ @EngSkills ➖
Funny Or Die (Youtube)
Drunk History vol. 6 (part 2) - Featuring John C. Reilly.
Subscribe now: https://www.youtube.com/c/funnyordie?sub_confirmation=1
Get more Funny Or Die
-------------------------------
Facebook: https://www.facebook.com/funnyordie
Twitter: https://twitter.com/funnyordie
Instagram: http://instagram.com/funnyordie
TikTok: funnyordie" rel="nofollow">https://www.tiktok.com/@funnyordie
➖ @EngSkills ➖
Language Log
AI Hyperauthorship
This paper's content is interesting — Mirzadeh, Iman, Keivan Alizadeh, Hooman Shahrokhi, Oncel Tuzel, Samy Bengio, and Mehrdad Farajtabar. "GSM-Symbolic: Understanding the Limitations of Mathematical Reasoning in Large Language Models." arXiv preprint arXiv:2410.05229 (2024). In short, the authors found that small changes in Grade-School Mathematics benchmark questions, like substituting different numerical values or adding irrelevant clauses, caused all the tested LLMs to do worse. You should read the whole thing for the details, to which I'll return another time.
But what inspired this post is a feature of that paper's bibliography, in which many items have a large number of authors. For example, this reference lists 65 authors before "and et al." [sic]:
Abhimanyu Dubey, Abhinav Jauhri, Abhinav Pandey, Abhishek Kadian, Ahmad Al-Dahle, Aiesha Letman, Akhil Mathur, Alan Schelten, Amy Yang, Angela Fan, Anirudh Goyal, Anthony Hartshorn, Aobo Yang, Archi Mitra, Archie Sravankumar, Artem Korenev, Arthur Hinsvark, Arun Rao, Aston Zhang, Aurélien Rodriguez, Austen Gregerson, Ava Spataru, Baptiste Rozière, Bethany Biron, Binh Tang, Bobbie Chern, Charlotte Caucheteux, Chaya Nayak, Chloe Bi, Chris Marra, Chris McConnell, Christian Keller, Christophe Touret, Chunyang Wu, Corinne Wong, Cristian Canton Ferrer, Cyrus Nikolaidis, Damien Allonsius, Daniel Song, Danielle Pintz, Danny Livshits, David Esiobu, Dhruv Choudhary, Dhruv Mahajan, Diego Garcia-Olano, Diego Perino, Dieuwke Hupkes, Egor Lakomkin, Ehab AlBadawy, Elina Lobanova, Emily Dinan, Eric Michael Smith, Filip Radenovic, Frank Zhang, Gabriel Synnaeve, Gabrielle Lee, Georgia Lewis Anderson, Graeme Nail, Grégoire Mialon, Guan Pang, Guillem Cucurell, Hailey Nguyen, Hannah Korevaar, Hu Xu, Hugo Touvron, and et al. The llama 3 herd of models. CoRR, abs/2407.21783, 2024. doi: 10.48550/ARXIV.2407.21783. URL https://doi.org/10.48550/arXiv.2407.21783.
Drilling down, that reference itself (The llama 3 herd of models") supplies its "contributor list" as an appendix:
http://languagelog.ldc.upenn.edu/myl/Llama3HerdTop.png
And the appendix splits the contributor list into two parts:
Llama 3 is the result of the work of a large number of people at Meta. Below, we list all core contributors (people who worked on Llama 3 for at least 2/3rd of the runtime of the project) and contributors (people who worked on Llama 3 for at least 1/5th of the runtime of the project). We list all contributors in alphabetical order of first name.
They then list 222 “Core Contributors” and 311 “Contributors”, for a total of 533 authors.
That's an order of magnitude smaller than (what I think is) the hyperauthorship record, the 5,154 authors for "Combined Measurement of the Higgs Boson Mass in pp Collisions at s= 7 and 8 TeV with the ATLAS and CMS Experiments", Physical review letters 2015.
➖ @EngSkills ➖
Phrasal Verb of the Day | Vocabulary | EnglishClub
take after
If you take after an older member of your family, you look like them or you have a similar personality to them.
➖ @EngSkills ➖
Word of the Day
whimsy
Definition: (noun) The trait of acting unpredictably and more from whim or caprice than from reason or judgment.
Synonyms: arbitrariness, flightiness, whimsicality, capriciousness.
Usage: She was an odd girl, prone to sudden flights of whimsy, and she floated through life cocooned in her own fantastic world.
Discuss
➖ @EngSkills ➖
Funny Or Die (Youtube)
Jon Gabrus Inside The FOD Vault: How Social Media Transformed Podcasting + Acting Nude With Al Roker
Actor, comedian, and podcaster Jon Gabrus (101 Places to Party Before You Die, High and Mighty, ActionBoyz) brings us “Gungan Style” from the Funny or Die Vault. He and our host, Marcos Gonzalez, take a deep dive into how the comedy landscape has changed since Jon got his start. From the internship to PA pipeline to TikTok dances and crazy stories on set, Jon and Marcos unpack where trend chasing is leading us.
Get notified when we drop new episodes, news, and show extras: https://norby.link/ctdAJD
Jon’s pick from the FOD Vault is “Gungan Style (Gangnam Style Parody)” from 2012: https://youtu.be/2df6x2YEeu8
Bio:
Jon Gabrus is an American actor and comedian, best known for his work on Guy Code, the podcast Comedy Bang! Bang!, and TVLand's Younger. He was a performer at the Upright Citizens Brigade Theatre and hosts the High and Mighty podcast (https://headgum.com/high-and-mighty), as well as co-hosting the Raised By TV and Action Boyz podcasts. To access a free episode to Jon’s podcast ActionBoyz, use the link: https://free.actionboyz.biz
0:00 Intro
3:17 Why Jon chose Gungan Style
6:50 College comedy group to PA pipeline
9:28 When video sketches started to take off at UCB
16:00 How podcasting has changed
25:23 The nature of trend chasing
36:38 Bombshell story, best and worst gigs
43:21 Drunk History story
Subscribe now: https://www.youtube.com/c/funnyordie?sub_confirmation=1
Get more Funny Or Die
-------------------------------
Facebook: https://www.facebook.com/funnyordie
Twitter: https://twitter.com/funnyordie
Instagram: http://instagram.com/funnyordie
TikTok: funnyordie" rel="nofollow">https://www.tiktok.com/@funnyordie
➖ @EngSkills ➖
Slang of the Day | Vocabulary | EnglishClub
gob (2)
to spit
➖ @EngSkills ➖
Idiom of the Day
a little leery
Cautious, wary, or uncertain of a given person, place, or thing. Watch the video
➖ @EngSkills ➖
Funny Or Die (Youtube)
Full Interview TOMORROW Jon Gabrus Acting Nude In Front Of Al Roker (Inside The FOD Vault Episode 3)
Subscribe now: https://www.youtube.com/c/funnyordie?sub_confirmation=1
Get more Funny Or Die
-------------------------------
Facebook: https://www.facebook.com/funnyordie
Twitter: https://twitter.com/funnyordie
Instagram: http://instagram.com/funnyordie
TikTok: funnyordie" rel="nofollow">https://www.tiktok.com/@funnyordie
➖ @EngSkills ➖
Funny Or Die (Youtube)
The Unexpected Celebrities Behind Drunk History's Origin Story
Drunk History creator Dreek Waters remembers how the famously booze-soaked history series came to be.
Check out the full episode of Derek Waters on Inside The Funny Or Die Vault: https://www.youtube.com/watch?v=t4TXV4PC2PA
Get all 10 episodes of season 1 now, and stay in touch for new episodes, news, and show extras: https://norby.link/ctdAJD
Subscribe now: https://www.youtube.com/c/funnyordie?sub_confirmation=1
Get more Funny Or Die
-------------------------------
Facebook: https://www.facebook.com/funnyordie
Twitter: https://twitter.com/funnyordie
Instagram: http://instagram.com/funnyordie
TikTok: funnyordie" rel="nofollow">https://www.tiktok.com/@funnyordie
➖ @EngSkills ➖
Language Log
Passed
There are many euphemisms for saying that someone died, two of the most common being "passed away" and "passed on". Lately, I've been hearing more and more people announce that so-and-so simply "passed". The first few times that I heard it spoken that way, I thought it sounded strange. Now, however, I'm so accustomed to this usage that it almost sounds normal, though I'm still barely to the point of being comfortable in saying it myself.
Substituting "passed" for "passed away" or "passed on" strikes me as being a euphemism for a euphemism.
Selected readings
* "Kim Jong Il: did he 'die' or 'pass away'?" (12/20/11)
* AIO:
Here are some words related to dying:
Synonyms: Perishing, succumbing, departing, expiring, disappearing, ending, fading, passing, deceasing, failing, dropping, kicking the bucket, biting the dust, conking out, consuming, drying up
Other words: Declining, disintegrating, ebbing, fated, final, mortal, sinking, vanishing, withering
Formal words: Fatality, casualty
Moribund: A word that means dying or in the process of passing from life
Different words can evoke different reactions in people. For example, "at peace" might sound more comforting than "He is dead". End of life workers should be aware that others may take offense at the choice of words.
In a western context, some people habitually say "passed away" or "passed" and are reluctant to say "died".
➖ @EngSkills ➖
Phrasal Verb of the Day | Vocabulary | EnglishClub
keep out
If you keep something out, you make sure it stays outside and doesn't come inside.
➖ @EngSkills ➖
Word of the Day
cogitation
Definition: (noun) Thoughtful consideration; meditation.
Synonyms: study.
Usage: After much cogitation he rejected the offer, deciding instead to pursue his dream of becoming an artist.
Discuss
➖ @EngSkills ➖
Language Log
Southeast Asians learning Mandarin
Anh Yeo is a Chinese from Vietnam. Currently she is studying in a graduate program of Chinese language and literature at Tsinghua University. To earn pocket money, she has taken up a job teaching Southeast Asia office workers Mandarin online. In response to this post "Aborted character simplification in the mid-1930s" (10/5/24), which had much to do with character simplification (or not) in Singapore, she wrote to me as follows:
I had two lessons tonight teaching Pinyin. Southeast Asians learn Pinyin fast (similar alphabet + existence of tones in Thai and Vietnamese), but because of that students are reliant on Pinyin and cannot remember characters! I have students learning for 3-4 months and still have to read off Pinyin (recognizing fewer than 50 characters). I always thought the coexistence of characters and Latin alphabet in Mandarin interesting!
What Anh said struck my fancy, and I wanted to learn more about her teaching experience, so I asked her these questions:
1. Are your students supposed to learn characters?
2. Do they try to learn characters? Just slow and difficult for them to do so?
3. Is their oral proficiency in Mandarin pretty good?
4. Do they do any reading and writing in Pinyin?
5. What is their ultimate goal in learning Mandarin? To use it in their office work?
Anh replied:
1/ The students are supposed to learn characters. The textbook I use to teach them is from the Princeton Language Program (which I also learned from 10 years ago), so the curriculum does stress learning characters (for all skillsets — reading, writing, listening, speaking). I do teach them characters, but given that the classes are online, it is hard to handwrite the characters for them to see, so I mostly type the characters using Pinyin. I feel like the biggest bottleneck is the way I input the characters so the students can see what they look like. Because all my classes are online, I cannot handwrite, so the students cannot remember the order of the strokes and prefer using pinyin to input characters for homework.
2/ My beginner students are mostly office workers (21 yo – 40 yo): busy and hardly have the time to handwrite the characters after class to practice/memorise. All of them asked for "quick tips" but most of them cannot commit the time and effort. Progress is very slow since it takes ~30 minutes every lesson for students to get familiar with the characters from the previous lesson again. Because the textbook has pinyin, they would read the pinyin instead of characters, so after 3 months, most of them are familiar with the "sound" but not the character.
3/ Their oral proficiency in Mandarin is very good. Because they are familiar with the concept of tones in their mother tongue, they can differentiate tones very well (and the pinyin alphabet is also very similar with the Vietnamese alphabet, most students only need 2-3 lessons to be able to pronounce Pinyin and tones).
4/ I feel like they do all of their reading and writing in Pinyin. Because the slides used in class have pinyin, they will mostly read the pinyin instead of the characters. I have tried to not include pinyin, but got complaints that without pinyin, group classes waste too much time in reading because students cannot recognise characters just yet. The slides I teach from have pinyin above all the characters (rearranged by WordPress to be in parentheses following each character), thus:
A：我(wǒ)今天(jīntiān)没有(méiyǒu)课(kè)，不(bù)忙(máng)。你(nǐ)现在(xiànzài)有(yǒu)空(kōng)吗(ma)？我(wǒ)请(qǐng)你(nǐ)喝(hē)咖啡(kāfēi)。
B：我(wǒ)不(bù)喝(hē)咖啡(kāfēi)，只(zhī)喝(hē)茶(chá)。
A：我(wǒ)有(yǒu)中国(zhōngguó)绿(lǜ)茶(chá)，也(yě)有(yǒu)英国(yīngguó)红(hóng)茶(chá)，你(nǐ)喝(hē)什么(shénme)茶(chá)？
B：红(hóng)茶(chá)、绿(lǜ)茶(chá)，我(wǒ)都(dōu)喝(hē)。
English translation added by VHM:
I don't have any classes today. I'm not busy. [...]
Phrasal Verb of the Day | Vocabulary | EnglishClub
hold down
to stop something from rising by pressing down on it or putting a heavy object on it
➖ @EngSkills ➖
Word of the Day
ostensible
Definition: (adjective) Appearing as such but not necessarily so.
Synonyms: seeming, apparent.
Usage: The ostensible reason of his appearance was the discovery, the very night before, of a "perfect little house."
Discuss
➖ @EngSkills ➖
Language Log
"Deppenapostrophe": Is English guilty after all?
Andreas Stolcke responds to "English is innocent" (10/10/2024):
The historical facts cited are correct, but they don't explain why the frequency of 's rose in the post-WW2 period, and again after about 2005 (= the internet), as indicated by the Google Ngram plot below.
The bump in the post-war era (after 1957) could be an effect of the Allied occupation (delayed by the book publishing process), which was reversed by the mid-1990s, and then encouraged again by the internet half a century later.
So my bet is still on an English (language) influence.
http://languagelog.ldc.upenn.edu/myl/GoogleDeppen1.png
However, it is true that use of genitive apostrophes was once much more pervasive, and dropped precipitously after around 1890, which is when Duden published his dictionary, which attained prescriptive (official) status in 1902. The mid-20th-century recovery was small in comparison to the earlier levels. All of this is nicely illustrated by Google Ngrams over the full time period:
http://languagelog.ldc.upenn.edu/myl/GoogleDeppen2.png
➖ @EngSkills ➖
Slang of the Day | Vocabulary | EnglishClub
intro
introduction
➖ @EngSkills ➖
Idiom of the Day
hit a brick wall
To come up against an immovable or insurmountable difficulty or obstacle. Watch the video
➖ @EngSkills ➖
Learn English Through Football Podcast: 2024-25: Spursy
Читать полностью…
