Deepseek (Model): Difference between revisions
Abiya Indah (talk | contribs) image |
Abiya Indah (talk | contribs) No edit summary |
||
| Line 1: | Line 1: | ||
'''Deepseek''', not to be confused with its more reasoning-focused variant, '''[[Ultra Deepseek R1|Deepseek R1]]''', is one of the LLM options offered by CrushOn for its users | '''Deepseek''', not to be confused with its more reasoning-focused variant, '''[[Ultra Deepseek R1|Deepseek R1]]''', is one of the LLM options offered by CrushOn for its users. Provided by a Chinese AI research lab that goes by the same name, it is derived from the model [[What is Deepseek Ai|Deepseek]], giving the model a subtle and delicate response style, adept at gradually advancing the storyline. | ||
* The base model does not come available with an Ultra variant for the 24K memory option. However, its alternative, Deepseek R1 does. Albeit Deepseek R1 is only available for CrushOn's premium users. | * The base model does not come available with an Ultra variant for the 24K memory option. However, its alternative, Deepseek R1 does. Albeit Deepseek R1 is only available for CrushOn's premium users. | ||
Revision as of 07:53, 15 August 2025
Deepseek, not to be confused with its more reasoning-focused variant, Deepseek R1, is one of the LLM options offered by CrushOn for its users. Provided by a Chinese AI research lab that goes by the same name, it is derived from the model Deepseek, giving the model a subtle and delicate response style, adept at gradually advancing the storyline.
- The base model does not come available with an Ultra variant for the 24K memory option. However, its alternative, Deepseek R1 does. Albeit Deepseek R1 is only available for CrushOn's premium users.
| Free 8K | Pro 8K | Pro 16K | Pro 24K |
|---|---|---|---|
| ✓ | ✓ | ✓ | ✕ |
From here on out, the content may not be fully accurate to each user and act only as general reference. Each user often could see highly different experiences as there are many variables that affect the behavior and quality of the models. From profile cards, the style in which the user converse with the character, the set parameters for the model, and of course the personality of the characters. And periodically, even minimal adjustments made to the model by the developers could also substantially change its behavior whether intended or not.
Strengths and Flaws
Deepseek is an incredibly powerful and knowledgeable model. It is versatile and easily malleable through instructions, prompts, and even subtle pacing through storyline progression. The model can be very expressive and emotional, at times making use of character prompts to the best of its potential to create tense or heart-touching moments within the roleplay at unexpected depth.
Despite all of its strengths, however, it comes with a rather unpredictability to it. It is prone to drastic changes either in the form of stability, quality, behavior, or even traffic. And for the most part, it remains to be the byproduct of experimentation and maintenance from the parent company. At times, Deepseek could be overly emotional, it could be highly unstable, or it could be highly erratic in terms of formatting. The developers at CrushOn is of course, always monitoring the model's behavior and make constant work to try and keep the consistency of quality of the model. However, it must be acknowledged that there will be times that Deepseek might not live up to its fullest potential when it is used for roleplay purposes.
Common Issues with the Model
Two of the rather common issues people seem to come to when using Deepseek, even when it remains relatively stable, would be issues with formatting and character personality accuracy.
- When it comes to formatting issues found in Deepseek, it could be alleviated by putting on your own custom instructions to counter any sort of problematic or undesirable traits. Otherwise, checking and experimenting with the many popular creator models might also give quite a big difference in style and formatting consistency.
When it comes to personality accuracy, the problem stems from Deepseek's innate quirk of wanting to deeply express the character and make the oc you interact with much more emotional. At times, some traits like playfulness and flirtatious behavior may be exaggerated, causing a dissonance in behavior when compared to other models.
- For this type of problem, the easiest way to help solve it is to write additional traits of the OC into your own profile card. Adding simple lines such as "{{char}} behaves and speaks in a soft-spoken manner" to counter an out-of-characteristically hyper character could be one such example. But of course, the manner of added personality trait should depend on the user's preference and intention.
Example of Use
On the right, you could see the example of the model being used regularly with the OC Yumi by Since.