Examples cited include whether or not a person has confirmed their e-mail address, if someone signs up for multiple accounts at the same time, behavior that could indicate a coordinated attack or accounts that repeatedly tweet or mention accounts that don't follow them.
According to Twitter, the majority of accounts reported make up less than 1 per cent of its total user base, which now stands at 336 million monthly active users, and it said much of what was reported didn't violate its rules.
"Some of these accounts and Tweets violate our policies, and, in those cases, we take action on them".
Twitter has made little effort to be transparent about the kind of signals it looks for when seeking to identify accounts that, in their words, "distort the conversation".
Today, we use policies, human review processes, and machine learning to help us determine how Tweets are organized and presented in communal places like conversations and search.
This latest improvement is part of a new initiative, announced by Twitter CEO Jack Dorsey in March, that aims to measure and improve the "collective health, openness, and civility of public conversation" on the platform.
Stormy Daniels' lawyer spotted with Christie Brinkley
She's claimed Trump hit on her while he was married to Ivana Trump , and when she was dating Billy. Trump to find someone else to fabricate things about me", the email advised.
Twitter will begin using a wider range of signals to rank tweets in conversations and searches, hiding more replies that are likely to be abusive, the company said today.
"There are many new signals we're taking in, most of which are not visible externally", said Del Harvey, vice president of trust and safety, and David Gasca, director of product management, health, in a blog post titled "Serving Healthy Conversation".
Twitter says it also looks at how accounts are connected to those that do violate rules and how they interact with each other. Hence, tweets that don't violate the rules but were submitted by users whom Twitter deems problematic will remain visible-you'll just have to click on "Show more replies" to access them.
But it adds: "Our work is far from done".
Still, Twitter's discourse is often colored by such tweets, and banning the people involved seems like a good way to open the platform to cries of censorship (well, more open than it is already). This technology and our team will learn over time and will make mistakes. Twitter also admitted that this is just one of several approaches meant to improve people's experiences on the platform, and that there will be "mistakes", "false positives", and "things we miss".