Code

[ENH] length controlled ALpacaEval#248

Code

Closed

YannDubs wants to merge 17 commits intomaintatsu-lab/alpaca_eval:mainfrom yann/length_correctiontatsu-lab/alpaca_eval:yann/length_correctionCopy head branch name to clipboard

+7,830-6,877

Commits

Commits on Mar 17, 2024

• [ENH] update codebase to use lenth corrected alpacaeval

  YannDubs

  committed
  b5cf7b4

• [ENH] update codebase to use lenth corrected alpacaeval

  YannDubs

  committed
  cad4a16

• scikit-learn insteall

  YannDubs

  committed
  607f3f6

• get_length_controlled_winrate only for AE2

  YannDubs

  committed
  2500c3b

• pass tests

  YannDubs

  committed
  9d43fd3

• pass tests

  YannDubs

  committed
  f9b8e3e

• pass doctests

  YannDubs

  committed
  2242cea

• update added models

  YannDubs

  committed
  019e255

• add claude and mistral annotators

  YannDubs

  committed
  af0601c

• update website

  YannDubs

  committed
  cee92e1

• update website

  YannDubs

  committed
  c376cd1

• add logging of extreme changes

  YannDubs

  committed
  7bf0f78

• [WIP] readme update

  YannDubs

  committed
  97b5aba

• [WIP] readme update

  YannDubs

  committed
  5fd41ce

• finish updating the readme

  YannDubs

  committed
  d296ed0

• push benchmark analysis

  YannDubs

  committed
  f15e44d

• merge new models

  YannDubs

  committed
  d0067d0