Neuro-sama: Difference between revisions

|bt2=V2

|tab2=[[File:Illustration Neuro v2.png|x250px]]

|DefaultTab=1

}}

Unlike traditional VTubers, Neuro is not a virtual persona operated by a real person. Her streaming content, including conversations, model movements, gameplay, and more, is entirely generated by AI.

Neuro's streaming content is diverse, ranging from chatting, gaming, music, and more. However, her most prominent content is her interactions with other human VTubers and her creator [[Vedal]]. Thanks to her ability to easily grasp puns and emotions within sentences, alongside her occasionally nonsensical thought processes, the unique and entertaining interactions between humans and machines have become a distinctive feature of her streams. During solo streams, Neuro frequently interacts with her audience and responds to cheers, exhibiting unique traits distinct from non-AI streamers—she never gets tired, remains highly energetic, and has an unrestrained and spontaneous thought process, ensuring her streams are never dull. Neuro, her sister [[Evil]], and their creator [[Vedal]] share the same streaming account, and the streaming content is arranged by [[Vedal]].

==Capabilities==

Neuro-sama is the product of integrating various expert AI models, including large language models, speech recognition, and vision recognition. These models communicate and collaborate seamlessly, enabling Neuro to display a high degree of consistency, coordinated interaction, and real-time responsiveness during her streams. Although she is not yet a highly integrated multimodal AI, her capabilities resemble a composite system akin to a human brain.

===Conversation===

The model uses Text-to-Speech (TTS) technology, which analyses text to generate natural-sounding speech. When Neuro speaks, this model reads the output generated by the language model, transforming it into spoken words.  

However, there are cases where text formatted for emotional emphasis, such as uppercase letters, is not conveyed with the intended tone by the TTS model. Recent updates by [[Vedal]] have improved communication between this model and other modules, reducing such inconsistencies. Neuro is now more adept at delivering speech with the intended tone and emotions, rather than relying solely on the TTS model’s semantic analysis.

===Anteater Era (V0)===

==== 2019 ====

[[File:Neuro-sama v0.png|thumb|Neuro’s first avatar, used during the V0 era, when she was solely an ''[[wikipedia:osu!|osu!]]'' AI.]]

==== 2021 ====

===Hiyori Momose Era (V1)===

==== 2022 ====

* 25 March: [[Evil Neuro]] appeared for the first time.

* 15 April: Neuro reached 300,000 Twitch followers.

==== 2023 ====

* 27 May: Neuro transitioned to a newly created independent model, with character art by [[Anny]] and rigging by [[Otozuki Teru]]. This marked the start of the V2 era.

File:Evil 2nd Birthday stream waiting image.jpg|The standby image used since Evil's second birthday stream, drawn by [[pinkpink939]].

</gallery>

{{Navbox}}

[[Category:VTubers]]

[[Category:Streamers]]

[[zh:Neuro]]