Because humans are imperfect actors, should the class of Basilisks include evidence in favor of hated beliefs?
It is unclear what will be the consequences and side-effects of not knowing the specific evidence. And on meta level: what will be the consequences of modifying your cognitive algorithms to avoid the paths that seem to lead to such evidence.
Depending on all these specific details, it may be good or bad. Human imperfection makes it impossible to evaluate. And actually not knowing the specific evidence makes it impossible again. So... the question is analogical to: "If I am too stupid to understand the question, should I answer 'yes', or should I answer 'no'?" (Meaning: yes = avoid the evidence, no = don't avoid the evidence.)
If it's worth saying, but not worth its own post (even in Discussion), then it goes here.