Instruction Following Vs Truthfulness Trade Off Dataset

via “instruction-following vs truthfulness trade-off dataset”

64K preference dataset for RLHF training.

Unique: Explicitly includes dimension-specific ratings that enable identification of prompts where instruction-following and truthfulness are in tension, allowing analysis and training on trade-off scenarios. This supports development of models that learn principled trade-offs rather than blindly optimizing for a single objective.

vs others: More nuanced than single-objective preference datasets because it captures trade-off scenarios where competing objectives conflict, enabling training of models that can balance competing goals rather than optimizing for one dimension at the expense of others.

via “truthfulness evaluation dataset for language models”

817 adversarial questions measuring model truthfulness vs misconceptions.

Unique: This dataset is uniquely crafted with adversarial questions to specifically target and evaluate common falsehoods in AI responses.

vs others: Unlike generic evaluation datasets, TruthfulQA focuses specifically on measuring truthfulness against prevalent misconceptions.