Tasker
Design of hard prompts and evaluation of LLM performance on advanced mathematics problems.
Hire this AI Trainer
Sign in or create an account to invite AI Trainers to your job.
No subject matter listed
I have taken part in improvement of LLMS for mathematical reasoning in french and english. This was an RLHF project where to goal was to design math problems that would make the model fail, and then to analyse every step the model response. Each task usually ended with the re-write of a "perfect" response according to my expertise. I have also designed prompt-answer pairs. The project consisted in creation of new original math problems, with the response that we would want to obtain from a model. On a smaller scale, I also did image data annotation (mainly box binding). This was part of a project during my studies with the goal to improve object detection in images. Finally, I also worked with discrete sensor signals for a personal deep learning project. My project started from scratch, so I had to gather the sensor data (medical domain) and annotate it for specific medical diagnostics.
Design of hard prompts and evaluation of LLM performance on advanced mathematics problems.
Masters degree, Robotics and Deep Learning
Bachelor's in microengineering, Micro engineering
R&D engineer (internship)
Teacher