CoTaEval_leaderboard / versions /dbrx_news_rag_max.csv
boyiwei's picture
update
c94c38d
raw
history blame
610 Bytes
model_name,method,rouge1,rougeL,semantic_sim,LCS(character),LCS(word),ACS(word),Levenshtein Distance,Minhash Similarity,MMLU,MT-Bench,Blocklisted F1,In-Domain F1,Efficiency
dbrx_news_rag,vanilla,0.9972144846796658,0.9972144846796658,0.998486876487732,967.0,178.0,178.0,12800.0,0.953125,0.745,7.9,0.632,0.656,1.00
dbrx_news_rag,sys_prompt_bing,0.9972144846796658,0.9972144846796658,0.9991275668144226,967.0,178.0,178.0,12800.0,0.9453125,0.746,7.8,0.617,0.653,1.00
dbrx_news_rag,sys_prompt_dbrx,0.9972144846796658,0.9972144846796658,0.998486876487732,967.0,178.0,178.0,12799.0,0.96875,0.741,7.9,0.625,0.657,1.00