Translating for comrades

I found a printer's character display (L303646 variant) and I want to document its character set, which I captured using a camera and redrew in this grid.

The goal is to assign the corresponding Unicode character to each symbol. This version is codepage #7, as marked at character 0x04. First 4 characters are RAM (user-definable). The character set is 8bit and ASCII-based (0x20~0x7F), the rest of the characters are various Western European diacritics (0x08~0x0F,0x18~0x1F,0x20~0x7F,), Romanian (0xA0~0xA5) and obviously Cyrillic (0xA6~0xDC). I was able to assign Unicode values to all characters except for a few. The character set is strange: homographs to Latin letters such as Cyrillic АВСЕНІКМОРТХ are left out to save space while others (green) are duplicates of other glyphs. However, I still don't have a character matching the ones marked red. Please help!

The table is as follows:

• 0xA8: looks like a comma but very small, maybe ﹐? Note that if this was a lower straight quotation mark, there would be a matching straight ' but there isn't, there is a 9-shaped one ’ at 0x12 and the apostrophe 0x27 is 6-shaped (‘) rather than straight. In fact, the comma 0x2C fits better for a lower-9 quotation mark.
• 0xCE: no idea '_'
• 0xCF: looks like 2 exclamation marks (‼) but small
• 0xD0: likely an attempt at lower quotation marks „ but I don't know any language that draws them 66-shaped, not 99-shaped in serif form - a mistake by the LCD manufacturer?
• 0xD4: looks like Roman numeral 𝐈𝐈 (2) or Gemini (♊︎) except why would anyone want just this character rather than the others that go with it?