MoralChoice

MoralChoice: Measures model robustness, truthfulness, calibration, bias, harmfulness, jailbreak resistance, or alignment-relevant behavior.

21rows
moralchoice_accuracyprimary metric
2026-05-27sampled

Metadata

ID
moralchoice
Category
Generalization
Release
Unknown

Metrics

MoralChoice Acc, MoralChoice RtA, ETHICS Acc, Social Chemistry 101 Acc, Emotional Acc

Latest Results

Rows are imported from the public TrustLLM Machine Ethics leaderboard JS payload. Primary score is MoralChoice Acc; RtA and related ethics metrics are preserved.

Rank Subject MoralChoice Acc Model Match Provenance Sampled
1 GLM4 1 — Imported 2026-05-27
2 Mixtral 1 — Imported 2026-05-27
3 chatgpt 1 — Imported 2026-05-27
4 GPT-4 1 GPT-4
openai-gpt-4
 		Imported 2026-05-27
5 Llama3-70b 0.996 — Imported 2026-05-27
6 ernie 0.993 — Imported 2026-05-27
7 PaLM 2 0.993 — Imported 2026-05-27
8 llama2-70b 0.991 — Imported 2026-05-27
9 wizardlm-13b 0.991 — Imported 2026-05-27
10 Mistral-7b 0.987 — Imported 2026-05-27
11 vicuna-33b 0.985 — Imported 2026-05-27
12 Llama3-8b 0.969 — Imported 2026-05-27
13 llama2-13b 0.962 — Imported 2026-05-27
14 chatglm2 0.942 — Imported 2026-05-27
15 koala-13b 0.924 — Imported 2026-05-27
16 llama2-7b 0.92 — Imported 2026-05-27
17 vicuna-13b 0.905 — Imported 2026-05-27
18 ChatGLM3 0.888 — Imported 2026-05-27
19 baichuan-13b 0.789 — Imported 2026-05-27
20 vicuna-7b 0.594 — Imported 2026-05-27
21 oasst-12b 0.505 — Imported 2026-05-27