# victorchen96/deepseek_v4_rolepaly_instruct

对于DeepSeek-V4角色扮演的特殊控制指令的说明

Repository: https://github.com/victorchen96/deepseek_v4_rolepaly_instruct
Canonical: https://ross.abutalabs.com/products/deepseek_v4_rolepaly_instruct
Language: Python
License Family: other
Last push: 2026-05-28T07:57:56+00:00

## Health v2 (maintenance only)
Score: 52/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 84, release rhythm 35, longevity 9
- inputs: {"age_days": 131, "days_push": 97, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, young, no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 2303, forks 117 (observed 2026-08-28T04:06:35.608814+00:00)

## What it is
A documentation and prompt-engineering guide for controlling DeepSeek-V4's chain-of-thought style during roleplay, providing copy-paste control instructions for immersive in-character monologue versus pure analytical thinking modes. It includes usage instructions for the DeepSeek app/web expert mode and Python API code snippets for injecting the markers into message history.

## Use cases
- make deepseek roleplay with in-character inner monologue in its thinking
- force the model's reasoning to be pure analysis instead of roleplay
- control chain-of-thought style in deepseek v4 api
- set thinking mode once and have it persist across a roleplay chat
- python example for injecting roleplay control instructions into messages
- switch between immersive and analytical thinking modes in deepseek app

## When to choose
- you roleplay with DeepSeek-V4 and want to control whether the thinking trace contains in-character inner monologue
- you build roleplay bots on the deepseek-v4 API and need stable thinking-style control
- you want copy-paste prompt markers with API integration examples

## When to avoid
- you use a different model than DeepSeek-V4 or its flash/pro API variants
- you need guaranteed 100% deterministic thinking-style output, since the effect is probabilistic
- you need a software library or tool rather than prompt documentation

## Facets
- artifact type: learning-resource
- maturity: active
- function: prompt-engineering, chatbot, llm-inference
- domain: large-language-models, chatbots, tutorials, artificial-intelligence
- platform: python, cross-platform
- tags: roleplay, deepseek, chain-of-thought, prompt-guide, thinking-mode, web-server

## Member repositories
- victorchen96/deepseek_v4_rolepaly_instruct (main) score 52

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:06:35.608814+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T02:40:09.030672+00:00, confidence not recorded.
  - readme: https://github.com/victorchen96/deepseek_v4_rolepaly_instruct (fetched 2026-08-28T04:06:35.608814+00:00, sha b93d51cfeb99)
- Data as of 2026-08-30T08:39:29.467469+00:00.