j6mes Profile Banner
James Thorne Profile
James Thorne

@j6mes

Followers
1K
Following
1K
Media
125
Statuses
764

CTO at Theia Insights. Assistant Prof at KAIST AI. PhD from @cambridge_cl. Co-organiser of fever workshop (https://t.co/9FTkHRwec2)

Cambridge, England
Joined October 2013
Don't wanna be here? Send us removal request.
@j6mes
James Thorne
5 days
RT @jiwoohong98: ⁉️Why do reward models suffer from over-optimization in RLHF?. We revisit how representations are learned during reward mo….
0
3
0
@j6mes
James Thorne
6 months
RT @vlachos_nlp: Pleased to announce the next @FEVERworkshop at ACL2025! Regular workshop papers (ARR and direct submissions) due 15th of A….
0
13
0
@j6mes
James Thorne
7 months
Looking forward to attending #NeurIPS2024 this week. Come meet @itsnamgyu who will present on block-transformer, which speeds up decoding by up to 20x.
@itsnamgyu
Namgyu Ho
1 year
Do you know your LLM uses less than 1% of your GPU at inference? Too much time is wasted on KV cache memory access ➡️ We tackle this with the 🎁 Block Transformer: a global-to-local architecture that speeds up decoding up to 20x 🚀. @kaist_ai @LG_AI_Research w/ @GoogleDeepMind 🧵
Tweet media one
2
4
25
@j6mes
James Thorne
8 months
NVIDIA's earnings call wasn’t about semiconductors — it was a vision for AI factories, 24/7 inference, and industry transformation. Using TIIC, we reveal $NVDA's diversification shows the future is broader than chips. What's your take? #AI #NVIDIA #LLMs
0
1
2
@j6mes
James Thorne
9 months
RT @RuscioValeria: 📢 Excited to share our new paper with @fabreetseo: "Beyond Position: How Rotary Embeddings Shape Representations and Mem….
0
29
0
@j6mes
James Thorne
1 year
RT @aliceoh: We are hosting wonderful NLP colleagues at KAIST on their way to ACL Bangkok! 🤩 On-site registration is closed, but the talks….
0
40
0
@j6mes
James Thorne
1 year
RT @RisingSayak: Aligning a diffusion model on preference data WITHOUT a reference model could be nice no?. So, @krasul and I are ideating….
0
13
0
@j6mes
James Thorne
1 year
RT @jiwoohong98: 🔥Mixtral-8x22B-base + ORPO🔥. 7k data & 1.3 hours to build a strong human-aligned 140B chat model🦾. 👉IFEval: 65%.👉BBH: 59%….
0
17
0
@j6mes
James Thorne
1 year
RT @aliceoh: Great work! I’m glad our benchmarks KoBBQ, KOLD, and CLIcK are used to evaluate HyperCLOVA_X. Making progress in Korean LLM!….
0
5
0
@j6mes
James Thorne
1 year
RT @dvilasuero: This is actually huge: . - No SFT stage (e.g., Zephyr used 200k examples).- Preference tuning with 7K examples only (other….
0
18
0
@j6mes
James Thorne
1 year
RT @TheiaInsights: 🚀Calling all astronaut economists! Introducing Theia Theme Connect, where you can travel the cosmos by connecting compan….
0
1
0
@j6mes
James Thorne
1 year
New preprint:.Align LLMs without reference models using ORPO ⬇️.Code and checkpoints available on @huggingface.
@jiwoohong98
Jiwoo Hong @ ICML2025🇨🇦
1 year
Align LLMs with the preference dataset ONLY with 💡ORPO💡. We introduce ORPO, alignment without reference model & SFT! With awesome dataset from @argilla_io + Mistral(7B) + ORPO, we present 🌟Mistral-ORPO-β🌟.🧵. 👉 AlpacaEval 2.0: 12.20%.👉 IFEval: 66.19%.👉 MT-Bench: 7.32
Tweet media one
Tweet media two
Tweet media three
0
0
6
@j6mes
James Thorne
1 year
RT @aliceoh: I can’t keep up with all these new papers and datasets from my students 😊 @j6mes and I wanted to give #kaist undergrad student….
0
2
0
@j6mes
James Thorne
1 year
Pleased to announce that Theia Insights has raised $6.5M to develop foundational AI technologies for the investment industry. Well done team!.
@TheiaInsights
Theia Insights
1 year
🚀Theia Insights is pleased to share news of our recent seed financing. We have secured $6.5M from some incredible investors to continue building #AI-based foundational technologies for the global #investment community - read more here:
Tweet media one
1
0
24
@j6mes
James Thorne
2 years
Congratulations Seonghyeon!.
@SeonghyeonYe
Seonghyeon Ye
2 years
FLASK🧪 is accepted to ICLR 2024 as Spotlight! Check out our work and see you in Vienna!.
0
0
3
@j6mes
James Thorne
2 years
RT @UKPLab: A group photo from the poster presentation of »AmbiFC: Fact-Checking Ambiguous Claims with Evidence«, co-authored by @Max_Glock….
0
6
0
@j6mes
James Thorne
2 years
RT @kaist_ai: ⭐️Congratulations to everyone at KAIST AI who have a total of 20 papers accepted to #NeurIPS2023! Check them here 🧵 https://t….
0
2
0
@j6mes
James Thorne
2 years
RT @kaist_ai: ⭐️Congratulations to everyone at KAIST AI who has a total of 11 papers accepted to #EMNLP2023! Check them here 📷🧵. https://t.….
0
8
0
@j6mes
James Thorne
2 years
RT @UKinKorea: Come and enjoy on Fri, Oct 27 at 6 pm, as @RMBandService and @ktigers_ent do a joint performance at 📍Gwanghwamun Plaza in fr….
0
3
0
@j6mes
James Thorne
2 years
Pleasure to host @OhadRubin at @kaist_ai to hear about his ongoing work of retrieval augmented language models, long range modeling and RPT
Tweet media one
0
4
35