Shivanshu Gupta
@shivanshug11
Followers
186
Following
378
Media
14
Statuses
82
PhD Candidate at UC Irvine | Research Intern @allen_ai | Previously @asapp @amazon @linkedin @msftresearch @iitdelhi | #NLP & #ML Research
Irvine, California
Joined April 2018
Happy to share that our GistScore paper has been accepted to #icml2024 -- the first ICML paper of my PhD! We introduced a simple yet state-of-the-art approach for selecting examples for in-context learning with LLMs. https://t.co/twYQ0hAer3 Checkout the thread👇for code/models.
🧵 Doing In-Context Learning with LLMs? Check out our GistScore paper for a simple approach to train example encoders that can be used to select informative examples for superior ICL performance. 💪 Or just use our multi-task trained encoders that work well out of the box! 🤗
3
2
16
🚀 Excited to present LAYERIF, our data-driven method for estimating layer quality in LLMs using Influence Functions! We show improvements in expert allocation (LoRA-MoE) + layer-wise pruning. Come meet us at #NeurIPS San Diego tomorrow 4:30pm! Paper link:
1
8
10
We’re Scaled Cognition, developing the first ever models trained specifically for agentic applications: 1. Our first system, APT-1, is now #1 on agentic benchmarks. 2. It was developed by a US team for a total cost of less than $11M. 3. Khosla Ventures led our seed round ($21M
9
43
237
Can AI really help with literature reviews? 🧐 Meet Ai2 ScholarQA, an experimental solution that allows you to ask questions that require multiple scientific papers to answer. It gives more in-depth, detailed, and contextual answers with table comparisons, expandable sections
14
73
221
📢 New paper from my internship at @cohere with @seraphinagt ‼️ Are you interested in investigating the fairness of LLMs in hiring contexts? Take a look at our work 🧵 https://t.co/zLy7TWLRAJ
arxiv.org
Large language models (LLMs) are increasingly being deployed in high-stakes applications like hiring, yet their potential for unfair decision-making remains understudied in generative and...
3
18
92
We (@peterjansen_ai, @mbodhisattwa, @tusharkhot, @harsh3vedi, @Hoper_Tom, @_DougDowney, @erichorvitz) are excited to announce the 📣1st Workshop on AI & Scientific Discovery Workshop (AISD), co-located with NAACL 2025. 📣 https://t.co/FtXgs19FKe
0
11
26
10 short videos about LLM infrastructure to help you appreciate Pages 12-18 of the DeepSeek-v3 paper ( https://t.co/OQv64u8cPP) 🧵 https://t.co/fYhnqxu813
12
122
732
Excited about #NeurIPS2024, my 15th one I think! Eager to meet everyone & hear abt your work! But if you want to hear me, there's an exciting panel tonight https://t.co/kgmOetZvYT Also @SpiffyAI is hiring ML engineers & @UCIbrenICS is hiring AI faculty, pls reach out to chat! 🧵
luma.com
From Research to Commercialization Join us for a conversation with speakers who made the leap from top research institutions to industry and are shaping how…
2
4
49
Excited to give an oral presentation of our work "Controllable Generation via Locally Constrained Resampling" @ #NeurIPS2024 SafeGenAI TL;DR We fix greedy constrained decoding using an ad hoc LLM approximation that we tractably condition on the constraint and reweighing samples
2
3
14
Had a great time at #EMNLP2024! Met lots of old friends and new. Having the entire #UCINLP group around made it one of the most fun conferences I have attended!
1
4
42
I will be presenting SUPER next week at EMNLP, Tuesday 4pm. Stop by to talk about evaluating agents on running research experiments and code in-the-wild!
📢 New Benchmark: SUPER for Setting UP and Executing tasks from Research repositories Reproducibility is crucial in science. We introduce SUPER to evaluate LLMs' capabilities in autonomously running experiments from research repositories. ⬇️ https://t.co/U47r3F3UO5
0
3
25
I am officially on the job market for industry research positions focused on agentic LLMs and multi-turn reasoning! I'll be at EMNLP next week and NeurIPS next month. Message me if you'd like to chat about jobs or LLM agent research. #EMNLP2024 #neurips2024 Personal links in🧵
8
21
82
I’ll be looking for summer research interns @allen_ai at the Aristo team (@ai2_aristo). Along with broad topics mentioned below, I’ll be focusing on model tuning and specialization for agents and data-driven discovery, in the paradigm of automated and AI-assisted scientific
11
37
303
Since I accidentally deleted the original tweet thread😅, this was work done during my summer internship at @asapp with my amazing mentors, Ethan @side1track1 and Clemens Rosenbaum!
0
0
2
What a great 1H it has been for 2025 for @USC_ISI NL Seminar 🎉, hosted by @cutelabname_nlp! We had an amazing line up of speakers that we were grateful for visiting ISI to share their inspiring work 🥰 If you have missed any of the talks, most of them are available on USC
1
1
13
Skill Set Optimization was accepted to @icmlconf 2024! I'm proud of this work and everything we learned about in-context policy improvement. Big thanks to my collaborators at @allen_ai. Way to go team!
Excited to share our work, "Skill Set Optimization", a continual learning method for LLM actors that: - Automatically extracts modular subgoals to use as skills - Reinforces skills using environment reward - Facilitates skill retrieval based on state https://t.co/vaSYjVzlB2 🧵
1
4
26
I'm happy to share that I have joined @allen_ai as a Research Intern with @ben_bogin and @tusharkhot! We're working on making LLMs capable of software engineering and research. Super excited! 😁 PS: I have moved to Seattle -- hit me up if you're around!
2
1
70
📢 This Thursday 11AM-12PM PST, we have @shivanshug11 from @UCIrvine giving us a talk at @USC_ISI on "Informative Example Selection for In-Context Learning." Join us to learn how to enhance in-context learning performance with better example selection!
1
1
13
Also, check out https://t.co/oZ3xoYZsIA for easy in-context learning experiments with a variety of datasets, LLMs, and example selectors!
github.com
Easy in-context learning experiemnts with variety of datasets, LLMs, and example selectors. - Shivanshu-Gupta/in-context-learning
0
0
1