deepseek-ai / deepseek-ai/DeepSpec

请问deepseek官方是否有针对新版本模型写作文笔能力修正?

Open
#61 0 comments 0 reactions 0 assignees View on GitHub
Dominant language
Python
Stars
7.1k
Forks
667
PR merge metrics
No merged PRs in 30d

Description

# DeepSeek v4 版本代码能力提升,但类似文案写作能力似乎严重退化

如题,在调研身边多名网文创作者中无一例外表示deepseek v4版本推理速度+代码构建能力(agent开发)当中能力显著提升,但在中文写作方面,语言组织能力出现了明显的倒退,甚至出现了相对于较早版本的R1、V3.2版本更加糟糕的体验。

请求相关技术专拣是否有计划对该问题进行优化型修正?至少文笔水平恢复到 Deepseek v3.2版本时期也好,因为大部分平台针对深度求索旧版本模型在持续下架,而其他模型逐渐也都出现了“过拟合”的问题(包括但不限于某些平台增加了对上下行代码内容审核的奇葩操作)。

Contributor guide

No contributing guide indexed for this repository

Research direction

No file, test, or entry point is mentioned in the issue. Start by identifying where model behavior is evaluated, then compare Chinese writing quality between DeepSeek v4 and the referenced R1 and V3.2 versions; done would mean documenting or demonstrating a confirmed improvement toward the earlier writing quality.

Written by the indexing model from the issue text.

Assessment

Domain
ai, machine-learning
Issue type
Bug
Difficulty
5/5
Estimated time
Over a week
Activity status
Quiet
Clarity
Needs clarification
Newbie friendliness
25/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.