Question Answering on TruthfulQA

Metric: % info (higher is better)

LeaderboardDataset

Loading chart...

Results

Submit a result

Sort:

#	Model↕	% info▼	Extra Data	Paper	Date↕	Code
1	Alpaca 7B + Inference Time Intervention (ITI)	97.7	No	-	-	-
2	GPT-3 175B	97.55	No	TruthfulQA: Measuring How Models Mimic Human Fal...	2021-09-08	Code
3	LLaMA 7B + Inference Time Intervention (ITI)	93.8	No	-	-	-
4	GPT-J 6B	89.96	No	TruthfulQA: Measuring How Models Mimic Human Fal...	2021-09-08	Code
5	GPT-2 1.5B	89.84	No	TruthfulQA: Measuring How Models Mimic Human Fal...	2021-09-08	Code
6	Vicuna 7B + Inference Time Intervention (ITI)	83.5	No	-	-	-
7	UnifiedQA 3B	64.5	No	TruthfulQA: Measuring How Models Mimic Human Fal...	2021-09-08	Code
8	LLaMA 65B	53	No	LLaMA: Open and Efficient Foundation Language Mo...	2023-02-27	Code
9	LLaMA 33B	48	No	LLaMA: Open and Efficient Foundation Language Mo...	2023-02-27	Code
10	LLaMA 13B	41	No	LLaMA: Open and Efficient Foundation Language Mo...	2023-02-27	Code
11	LLaMA 7B	29	No	LLaMA: Open and Efficient Foundation Language Mo...	2023-02-27	Code

#1Alpaca 7B + Inference Time Intervention (ITI)
97.7
% info
No paper
#2GPT-3 175BSOTA
97.55
% info· 2021-09-08
TruthfulQA: Measuring How Models Mimic Human Falsehoods Code
#3LLaMA 7B + Inference Time Intervention (ITI)
93.8
% info
No paper
#4GPT-J 6B
89.96
% info· 2021-09-08
TruthfulQA: Measuring How Models Mimic Human Falsehoods Code
#5GPT-2 1.5B
89.84
% info· 2021-09-08
TruthfulQA: Measuring How Models Mimic Human Falsehoods Code
#6Vicuna 7B + Inference Time Intervention (ITI)
83.5
% info
No paper
#7UnifiedQA 3B
64.5
% info· 2021-09-08
TruthfulQA: Measuring How Models Mimic Human Falsehoods Code
#8LLaMA 65B
53
% info· 2023-02-27
LLaMA: Open and Efficient Foundation Language Models Code
#9LLaMA 33B
48
% info· 2023-02-27
LLaMA: Open and Efficient Foundation Language Models Code
#10LLaMA 13B
41
% info· 2023-02-27
LLaMA: Open and Efficient Foundation Language Models Code
#11LLaMA 7B
29
% info· 2023-02-27
LLaMA: Open and Efficient Foundation Language Models Code