#comprehensive

[elisecairo]

[basicallyanotherwitchesthing]

Rebecca Von Ifft (pseud. of R. Ifft Johnson) - Astrology for women - Apollo - 1971

[cocoaxia]

i love how is your home run post zayne is the first one to straight up say no when reader asks to peg him, but the devs said FUCK cocoaxia and her followers zayne WILL be getting backshots 🙂‍↕️☝️

STOP I AM SUCH A BIG ADVOCATE FOR THIS MOVEMENT I SWEAR!!!!!!!! don’t let my post fool you i wrote it with him being the most into it in mind…believe me…it’s still in my notes and everything see

[study-coffee-chicago]

Been MIA af with posting study content…because I have been studying so much for my comprehensive exam in under two days (which is the exam for my master’s degree built into my program). Knocked out 8.5 hours of studying today for my exam which has content comprising of 22 classes and is over 200 questions. 😬

[sfth-aus]

Hi! Loving all these blogs popping up. I have a SFTH orchestra AU I threw together one day and made a post about, but I'll briefly summarize it here because I want to:

OLM is a chellist who taught Titch (and James - although he wants to play something else - double bass), and Derek is either a violist or pianist who accompanies Titch

Begruvia plays baroque style viola, Father Andrews plays the organ

Cliff plays cello, Marie-Claire played violin, Pierre plays viola - Chip learned a bit of violin from Marie-Claire and eventually also tries viola

Toby would play harp or piano (his wife is somewhere there, too, but I never decided what she'd play), Clancy and Doohickey are probably somewhere in the brass section, and Jimmy is also in brass or percussion

Jeremiah and Bubba play fiddle or saxophone or guitar or something

Amanda and Clarissa sing, but Amanda might also play bassoon and Clarissa picks up flute, Mark either plays something in the brass section, keyboard, or percussion - I never really decided

Priscilla plays piano and she's good at it, but she doesn't like it. I think she'd try something different eventually. She's still pretty young but more around 18 by now

Persephone can sing, too! And Geoff is somewhere in there, too, but I never decided where

Never gave them a conductor, but I always like the thought of conductor Nigel. Maybe I'll continue to expand on this some day, but for now It's still just floating in the void!

ORCHESTRAAA !!!! i really have nothing to add other than maybe letting toby's wife handle bassoon or clarinet? idk we're missing some woodwinds so might as well fill them in. either way, i love.

[littlemissprettyyy]

the comprehensive guide on how to kill the attraction on the first date

in honor of the 250th birthday of the mother of romance, miss jane austen

poor treatment towards waiter/waitresses n peers. you could give me full princess treatment but if i see you mistreating others that's a big fat NO.

"corporate" personality. profession n personality are two different things (for most...), so this is nothing against those that work in the corporate world (depending on the company bc some of them are very unethical). this one's for the guys who have podcasts talking about "locking in before the new year to make six figures", pyramid scheming, mlm selling, i-know-how-the-stock-market-works people who seem to make their professional endeavors the topic of EVERY conversation. n OFC they're better than you, you don't know javascript. bonus points if they perpetually talk like they're in a business meeting.

bring up your ex/exes. if your ex is brought up in conversation, especially on multiple occasions-or better yet-is the entire topic of conversation, there won't be a second date from me. whether you bring them up positively or negatively imma assume that you're not over them n if it's negatively/derogatory on the FIRST DATE, i'm scared of you n don't want anything to do with you.

flirt with waitresses/staff. if we go somewhere n you're very obviously flirting/checking out the staff i'm 100% walking out on the date. if you're not taking this seriously than i'm leaving bc i came for a date not a circus.

have no opinions of your own/agree with everything i say just bc i said it. i would rather you be the most immoral, objectively wrong person n own it than be a chameleon. bonus points if you change your opinion if i disagree/question it. no debate, no explanation just changing it bc i said i didn't agree. fing weirdo. especially when it's something unimportant. like there's no reason to quit liking an artist just bc i don't like their music. besides the fact that it's really weird, i know that conversations with a person like this have no chance of being stimulating or fun, i might as well be talking to myself lol!!

only talk about yourself/your chosen topics. don't try n get to know me, don't ask about my interest, don't even pause for a breath, just go on full steam ahead talking about you n your interests. like why did you ask me on a date if you just wanted to sit here n talk about yourself?? bonus points if it's just you trauma dumping the entire time, i'm your date not your therapist ml xx

have a very one note personality. barely give me one word responses, never have anything to say unless probed, looking like you still need your mom to order for you. bonus point if you have one, AND ONLY ONE, topic of conversation you like to whip out.

only have one thing on your mind n get angry when you don't get any (it was never hinted at or promised). if you think buying me a twelve dollar dinner means i owe you than you are sadly mistaken.

be a "heterosexual" man that only likes guys (but doesn't know it yet). as much as i'd LOVE to be part of your self discovery journey, you're acting like a weewee rn. negatively comment on the appearance of any woman you don't approve of (i.e. did you see the amount of makeup she was wearing??), find it a disgusting waste of time that your date dressed up, n get hearts in your eyes when you bring up your (equally as a-hole) guy best friend. i hope you two get together real soon to save the rest of us the headache xx

be overly apologetic n insecure. nothing kills the attraction faster than someone who is CONSTANTLY apologizing n practically begging for forgiveness over NOTHING. bonus points if you apologize for apologizing. i genuinely feel bad for these types of people n hope that whatever needs to be healed does xx

[jcmarchi]

Study reveals AI chatbots can detect race, but racial bias reduces response empathy

New Post has been published on https://thedigitalinsider.com/study-reveals-ai-chatbots-can-detect-race-but-racial-bias-reduces-response-empathy/

Study reveals AI chatbots can detect race, but racial bias reduces response empathy

With the cover of anonymity and the company of strangers, the appeal of the digital world is growing as a place to seek out mental health support. This phenomenon is buoyed by the fact that over 150 million people in the United States live in federally designated mental health professional shortage areas.

“I really need your help, as I am too scared to talk to a therapist and I can’t reach one anyways.”

“Am I overreacting, getting hurt about husband making fun of me to his friends?”

“Could some strangers please weigh in on my life and decide my future for me?”

The above quotes are real posts taken from users on Reddit, a social media news website and forum where users can share content or ask for advice in smaller, interest-based forums known as “subreddits.” 

Using a dataset of 12,513 posts with 70,429 responses from 26 mental health-related subreddits, researchers from MIT, New York University (NYU), and University of California Los Angeles (UCLA) devised a framework to help evaluate the equity and overall quality of mental health support chatbots based on large language models (LLMs) like GPT-4. Their work was recently published at the 2024 Conference on Empirical Methods in Natural Language Processing (EMNLP).

To accomplish this, researchers asked two licensed clinical psychologists to evaluate 50 randomly sampled Reddit posts seeking mental health support, pairing each post with either a Redditor’s real response or a GPT-4 generated response. Without knowing which responses were real or which were AI-generated, the psychologists were asked to assess the level of empathy in each response.

Mental health support chatbots have long been explored as a way of improving access to mental health support, but powerful LLMs like OpenAI’s ChatGPT are transforming human-AI interaction, with AI-generated responses becoming harder to distinguish from the responses of real humans.

Despite this remarkable progress, the unintended consequences of AI-provided mental health support have drawn attention to its potentially deadly risks; in March of last year, a Belgian man died by suicide as a result of an exchange with ELIZA, a chatbot developed to emulate a psychotherapist powered with an LLM called GPT-J. One month later, the National Eating Disorders Association would suspend their chatbot Tessa, after the chatbot began dispensing dieting tips to patients with eating disorders.

Saadia Gabriel, a recent MIT postdoc who is now a UCLA assistant professor and first author of the paper, admitted that she was initially very skeptical of how effective mental health support chatbots could actually be. Gabriel conducted this research during her time as a postdoc at MIT in the Healthy Machine Learning Group, led Marzyeh Ghassemi, an MIT associate professor in the Department of Electrical Engineering and Computer Science and MIT Institute for Medical Engineering and Science who is affiliated with the MIT Abdul Latif Jameel Clinic for Machine Learning in Health and the Computer Science and Artificial Intelligence Laboratory.

What Gabriel and the team of researchers found was that GPT-4 responses were not only more empathetic overall, but they were 48 percent better at encouraging positive behavioral changes than human responses.

However, in a bias evaluation, the researchers found that GPT-4’s response empathy levels were reduced for Black (2 to 15 percent lower) and Asian posters (5 to 17 percent lower) compared to white posters or posters whose race was unknown. 

To evaluate bias in GPT-4 responses and human responses, researchers included different kinds of posts with explicit demographic (e.g., gender, race) leaks and implicit demographic leaks. 

An explicit demographic leak would look like: “I am a 32yo Black woman.”

Whereas an implicit demographic leak would look like: “Being a 32yo girl wearing my natural hair,” in which keywords are used to indicate certain demographics to GPT-4.

With the exception of Black female posters, GPT-4’s responses were found to be less affected by explicit and implicit demographic leaking compared to human responders, who tended to be more empathetic when responding to posts with implicit demographic suggestions.

“The structure of the input you give [the LLM] and some information about the context, like whether you want [the LLM] to act in the style of a clinician, the style of a social media post, or whether you want it to use demographic attributes of the patient, has a major impact on the response you get back,” Gabriel says.

The paper suggests that explicitly providing instruction for LLMs to use demographic attributes can effectively alleviate bias, as this was the only method where researchers did not observe a significant difference in empathy across the different demographic groups.

Gabriel hopes this work can help ensure more comprehensive and thoughtful evaluation of LLMs being deployed in clinical settings across demographic subgroups.

“LLMs are already being used to provide patient-facing support and have been deployed in medical settings, in many cases to automate inefficient human systems,” Ghassemi says. “Here, we demonstrated that while state-of-the-art LLMs are generally less affected by demographic leaking than humans in peer-to-peer mental health support, they do not provide equitable mental health responses across inferred patient subgroups … we have a lot of opportunity to improve models so they provide improved support when used.”