Nina

Studying LLM reasoning, safety & behavior | MSc student @KSU_CCIS

[Nina @_NeuralNerd]

🚀Introducing ANERR: Arabic NLP Essential Resource Repository

Curated 220+ Arabic NLP papers, datasets, surveys, tools, organized by task, year & topic

مستودع أَنِير: مرجع بحثي منظم لمصادر معالجة اللغة العربية!

https://t.co/zBLzrDlpqa

✨Contributions welcome!
#NLP #ArabicNLP

[Nina @_NeuralNerd]

A thought that stayed with me after studying privacy attacks in my AI ethics course:

> Open-weight models are not just trying to catch up with closed models.

> They also face a different threat model; once people can run them locally, fine-tune them, or modify them, evaluation becomes harder.

As someone excited about contributing to open-source LLMs, I think capability alone is not enough. We also need to ask how well these models hold up under open access.

[Nina @_NeuralNerd]

I did not expect one book to carry me through this much of the semester, but Hands-On Large Language Models genuinely helped 😭

It gave me enough foundation to compete in an NLP shared task, understand LLM papers better, and even prepare a lecture presentation for my data mining course 🥹

[Jawahar ☕ @codevsdev]

Which book, course, or resource changed your career?

[Nina @_NeuralNerd]

Update: Less chaos did not happen 💀

But the semester had its moments:
- 2nd place in an Arabic NLP competition
- one paper accepted, another still under review for 2 months ... :)
- my first Arduino project with questionable cable management
- aquascaping experiments
- my cat fighting FIP like a little warrior (3 months and still strong 🤍🙏)

Finals are done, and next semester is the master's project 🌱

[Nina @_NeuralNerd]

Gm 🌱
First day of 2nd semester.
Hoping for less chaos and more focus on code and experiments.

[Nina @_NeuralNerd]

@harsh_rawa47363 ✨️ thanks

[Nina @_NeuralNerd]

Gm 🌱
First day of 2nd semester.
Hoping for less chaos and more focus on code and experiments.

[Nina @_NeuralNerd]

@harsh_rawa47363 2 course 😬 datawarehouse and Ai ethics.

[Nina @_NeuralNerd]

@BhargavLearns Thanks Bhargav!

[Nina @_NeuralNerd]

[Day 24 of LLM Grind] 🤗

~ 3 hours of coding
> Deep dive into Arabic text normalization, handling diacritics, elongations, emojis, and noise.

> Explored feature fusion
Planning a multi-feature setup (text + metadata) for improved classification.

[Nina @_NeuralNerd]

Day 23 of LLM Grind ✨
Lazy day today, not feeling 100%, so I kept it light.

~2 hours of chill studying
> Completed a Kaggle certificate in AI Ethics ✅️

Not every day has to be intense.
Consistency > intensity 😌.

[Nina @_NeuralNerd]

@BhargavLearns Welcome back 👍

[Nina @_NeuralNerd]

@BhargavLearns I only understood 1 word, and that's diabetes for some reason 😂😭 but less go!!!

[Nina @_NeuralNerd]

@BhargavLearns Masters ⚔️

[Nina @_NeuralNerd]

@AviiWeb3 Exactly 💯

[Nina @_NeuralNerd]

@UnpaidAvenger

[Nina @_NeuralNerd]

@herooffjustice @Abde_Mustafa0 Niceee, maybe I should try that as well 🤔

[Nina @_NeuralNerd]

@Adidotdev Good luck Adit ✨️

[Nina @_NeuralNerd]

@ibaame بتوفيق في رحلتك + جهازك نفس جهازي 😂

[Nina @_NeuralNerd]

"What's reality? I don't know. When my bird was looking at my computer monitor I thought, 'That bird has no idea what he's looking at.' And yet what does the bird do? Does he panic? No, he can't really panic, he just does the best he can. Is he able to live in a world where he's so ignorant? Well, he doesn't really have a choice. The bird is okay even though he doesn't understand the world. You're that bird looking at the monitor, and you're thinking to yourself, 'I can figure this out.' Maybe you have some bird ideas. Maybe that's the best you can do." ~ Terry Davis

[Nina @_NeuralNerd]

Day 23 of LLM Grind ✨
Lazy day today, not feeling 100%, so I kept it light.

~2 hours of chill studying
> Completed a Kaggle certificate in AI Ethics ✅️

Not every day has to be intense.
Consistency > intensity 😌.

[Nina @_NeuralNerd]

Day 22 of LLM Grind

~ 4 hours study session
> Changed my visualization library from seaborn to plotly 🤗 all because I needed to plot emojis!
> Did minor text preprocessing (not sure how to go about tbh)

Trends observed in dataset:
- Most polite posts are coming from YouTube.
- Most impolite posts are directed at companies.
- Top 5 emojis used in polite posts: [❤️, 🤍, 👌,🌹,🩶]
- Top 5 emojis used in impolite posts: [💔,👇,😂,🔥,😭]
- There are not that many posts with emojis but most posts with emojis are polite.
- Posts contains links and tags.
- Posts that do contain links are mostly neutral or polite, a small portion is impolite.

[Nina @_NeuralNerd]

@amartyaintech Gm ✨️

[Nina @_NeuralNerd]

@Abde_Mustafa0 Thanks 🔥

[Nina @_NeuralNerd]

Day 22 of LLM Grind

~ 4 hours study session
> Changed my visualization library from seaborn to plotly 🤗 all because I needed to plot emojis!
> Did minor text preprocessing (not sure how to go about tbh)

Trends observed in dataset:
- Most polite posts are coming from YouTube.
- Most impolite posts are directed at companies.
- Top 5 emojis used in polite posts: [❤️, 🤍, 👌,🌹,🩶]
- Top 5 emojis used in impolite posts: [💔,👇,😂,🔥,😭]
- There are not that many posts with emojis but most posts with emojis are polite.
- Posts contains links and tags.
- Posts that do contain links are mostly neutral or polite, a small portion is impolite.

[Nina @_NeuralNerd]

[Day 21 of LLM Grind]
- Spent ~8 hours today working on Arabic politeness classification of social media posts 🧪

- Did some dataset exploration and modeling choices.

#NLP #LLM #BuildInPublic