Seungone Kim

icon indicating a website link seungonekim.github.io

icon company Carnegie Mellon University icon location Pittsburgh, US icon location Ph.D. student at CMU LTI working on LLM Evaluation. Prev. KAIST AI, Yonsei Univ, NAVER AI Lab, LG AI Research #nlproc

Results 5 comments of Seungone Kim

multi gpu inference with run_rm.py

Thanks for your response @natolambert! I was trying to test generative reward modeling (with GPT-4, Prometheus, Auto-J) and it seems like `run_dpo.py` has a slightly different functionality than what I...

Consider our CoT dataset (CoT Collection)

Sounds fair enough!! I'll organize the code & upload the models until this week. Thanks in advance!

Using Language Models as Evaluators

@lintangsutawika Is there currently any initiative for this feature? I would love to help

Using Language Models as Evaluators

@baberabb Hello Baber, nice to meet you! I'd love to collaborate with you on working on this! On which platform could I best communicate with you (e.g., slack, discord)?

Hi. Is it support to korean language?

Hello @95jinchul , the model isn't specifically trained to evaluate Korean responses, but it does work to some extent. The trick here is to only convert the {instruction, response} into...