A gaming platform's Foundry-hosted assistant helps players discuss in-game strategy. Reviewers notice two problems: the assistant occasionally repeats a slang term that is specific to this game's toxic community and not recognized by any default content category, and separately, it is over-blocking harmless in-game combat descriptions that use words like 'kill' and 'attack' in a clearly fictional gameplay context. (Select TWO.) Which two configuration changes address these problems?
Select all correct answers, then click Submit.
Short Explanation
Two separate problems need two separate fixes here. The slang term isn't hateful or violent by any standard definition, so no severity dial will ever catch it — you need to name it explicitly on a custom list so it gets flagged no matter what category it doesn't fit into. The over-blocking issue is the opposite kind of problem: the system is treating mild, clearly fictional combat talk as if it were as risky as genuinely harmful violent content, so you loosen how sensitive that one category is, letting only the higher-severity stuff get stopped. Turning the whole violence filter off would fix the annoyance but also tear away protection against content that actually deserves blocking, which is way more than this situation calls for. And fiddling with how creative or random the model's wording is won't touch either problem — that setting has nothing to do with banned terms or filter sensitivity.
Full Explanation
The correct answers are B and C. Adding the game-specific slang term to a custom blocklist lets the content safety system flag that exact term even though it does not fall under any default harm category, which directly solves the first problem. Raising the severity threshold for the violence category means only higher-severity content gets blocked, so low-severity, clearly fictional combat language like 'kill' and 'attack' in a gameplay context passes through instead of being over-blocked, which solves the second problem. Option D is incorrect because disabling the violence filter entirely removes protection against genuinely harmful violent content along with the fictional gameplay language, which is a far blunter change than the team needs and increases risk elsewhere. Option A is incorrect because temperature controls how varied or creative the model's word choices are; it has no mechanism for suppressing a specific banned term or for changing how the content safety system classifies severity, so it would not reliably stop the slang term from appearing or fix the over-blocking issue.