Recently, I've done a few focused reviews with just the "Kana→meaning" vector enabled (normal reviews have Japanese→meaning, and so usually Kanji→meaning); for words with kanji, it feels like a whether a question will contain kanji with full furigana, or kanji withno furigana at all is determined by a coin toss. Regardless, answering correctly seems to increase the Kana→meaning vector mastery (which seems silly when a question might look like what is pictured in the screenshot). I'm not sure whether kanji+full furigana for the quizzed term or no kanji, only kana would be the better option here.
By chance, do you have the experimental setting for furigana turned on? I cannot think how that would interfere, but it is worth removing that as a potential cause before moving forward.
I do have "Show furigana in answers" enabled (along with "Show furigana for words I haven't studied"). I tried disabling "Show furigana in answers" just now, and the issue persisted anyway (i.e. there were still questions with both kanji+furigana and kanji+no furigana for the quizzed term in a Kana→meaning focused review).
Thanks. I was not able to replicate it, but I have a suspicion I know what is causing it. I made a few changes, so please give it a try and let me know if you notice an improvement.
In today's kana→meaning focused review, I got 首/くび with furigana, and 申し上げます without furigana in sentence-based questions, and 鼠講 without furigana and sentence. All of those kanji are known, so I'm not sure what made 首 special there.
