Phase 1
The first phase of our operation focuses on fostering active participation from users as they engage with Large Language Models (LLMs). This phase is pivotal as it lays the groundwork for the further development of our platform and the overall optimization of the LLMs. We encourage users to provide a holistic range of feedback on the LLM's responses. This can range from a simple scoring scale based on their level of satisfaction, to more nuanced insights into the areas where the LLM excels and where it could be improved. This comprehensive feedback process is designed to capture a broad spectrum of user interactions and preferences. The resulting dataset is invaluable for the ongoing refinement of our preference and reward models, and ultimately for enhancing the performance and user experience of partner LLMs.
Targets for the first phase of our operations are as follows:
Launch of Public Sale via Liquidity Bootstrapping Pool: Capital raised via our LBP will be utilised for further development of our products, operations and growth and team expansion plans.
Token Generation Event and Liquidity Lockup: Following our LBP sale, we will be initiating the token generation event. Following the generation event, implementing a liquidity lockup to ensure deep liquidity of our token for holders.
Onboarding Partner LLMs and Iterative Retraining via RLHF: A paramount goal for us in this phase is to onboard a diverse range of partner LLMs. Once onboarded, we will carry out multiple iterations of retraining via our Reinforcement Learning from Human Feedback (RLHF) system. This iterative process will not only enhance the LLMs' performance but also provide us with a wealth of data that will be instrumental in reducing the time for each training iteration.
Enhancements to the HapticScore System: In this phase, we will also be focusing on iterating and improving our HapticScore system. This system plays a vital role in rating and evaluating the feedback received from users. By enhancing this system, we can ensure that the feedback is accurately assessed and effectively utilized for the continuous improvement of our partner LLMs.
Launch of the First Incentive Program: As part of our efforts to encourage user participation and engagement, we will be launching our first incentive program. This program will reward users for their valuable contributions in terms of feedback and will play a key role in fostering a vibrant and active community.
In conclusion, our first phase encompasses a host of vital objectives and milestones that are integral to our journey towards revolutionizing the way LLMs are trained and optimized.
Last updated