These are so close with regard to how recent events are classified.
Let’s say someone has a tendency to recall recent events more vividly, and gives more undue weight to that recent event.
This is an availability and representative bias. So what is the differentiation?
In your example, I would lean towards representative bias because it implies that past information will likely continue going forward which is a cognitive bias, specifically a cognitive error: belief perseverance. Here are the definitions.
Availability Bias just puts an unnecessary emphasis on information that’s easily/readily available. ex) 2019 CFA Level III Candidates practice by taking 2017 and 2018 mock exams…because they’re readily available.
Representation Bias is just the idea that the past will persist and new information is based on past experience ( base rate neglect : messing up the initial ‘base case’ classification and using that initial classification as 100% correct with no consideration that it could actually have been wrong from the get go. There’s also sample-size neglect which makes the initial classification on an overly small sample size like making a decision to purchase a stock solely based on dividend yield without looking at any other characteristics of value stocks).
Maybe these will help, but I’ve come up with a few mnemonics to differentiate between cognitive and emotional biases and their sub-groups:
"If you BELIEVE you made an ERROR, go C.C.C.H.R"
Cognitive Errors: Belief Perseverance:
C onfirmation, C ontrol, C onservatism, H indsight, R epresentative Bias
"When processing information, make sure to THINK like F.A.M.A!"
Cognitive Errors: Information Processing:
F raming, A nchoring/ A djustment, M ental Accounting, A vailability Bias
"Don’t be such EMOTIONAL L.O.S.S.E.R.S!"
Emotional Biases:
L oss-Aversion, O verconfidence, S elf-attribution, S elf-control, E ndowment, R egret aversion, S tatus Quo Bias
Hope this helps 
I like your acronyms!
From definition of representativeness bias, “past will persist” means overweighting past experience.
But in base rate neglect, it turns out to be underweighting past experience, instead, decision is made exclusively based on new information.
I think they are contradictory. The original definition of representativeness bias looks more like a conservatism bias. The base rate neglect just shows an opposite scenario.