ezelikman/quietstar-8-ahead

Model

stars

commits

repos using this model

linked in READMEs

Mar 23, 2024

updated

endpoints_compatible

mistral

safetensors

text-generation

text-generation-inference

transformers

README

Mistral-7b with continued pretraining using Quiet-STaR (https://arxiv.org/abs/2403.09629) for generating 8 thought tokens before each output token.

Contributors

ezelikman

4 commits

Similar repos

ezelikman/quiet-star

Code for Quiet-STaR

739

yu-zh47/Adaptive_QuietSTaR

google/gemma-7b

Model

3,406

epfl-llm/meditron-7b

Model

327

omni-research/Tarsier-7b

Model

PygmalionAI/Eleusis-12B-FP8

Model

google/gemma-7b-it

Model

1,250

google/gemma-2-27b

Model

211

Used in code

ostix360/Quiet_STaR

This project aims to implements quiet_star algoithm

2 files

yu-zh47/Adaptive_QuietSTaR

2 files

marwanabdelhameed/Reasoning-vs-InfernceScaling

Research Paper (https://arxiv.org/pdf/2412.16260)

2 files

ezelikman/quiet-star

Code for Quiet-STaR

739

1 file

wise-east/ntp_rl

reinforcement learning with next token prediction as the reward

1 file

Linked in READMEs

ezelikman/quiet-star

Code for Quiet-STaR

739

README badge

yu-zh47/Adaptive_QuietSTaR

README badge