专注A股市场每日实战操作，让2亿散户不迷路！如果你不是长线投资者，想在A股市场中活下来，那么必须明白，周密的布局计划只能有个好的开始，严格的操盘纪律才能结出丰硕的果实！

　　周四大盘窄幅震荡，稳定在20日线和3300整数关之上。市场运行趋势总体还算平稳，但是上行节奏并不快，尚未摆脱当前台阶进入到下一个台阶中去。

　　从盘面上看，沪深两市个股跌多涨少，亏钱效应明显。从板块涨跌幅上看，ML Ops概念领涨，半导体产业链、贵金属等涨幅居前；ChatGPT概念领跌，数字经济主题跌幅居前。近期主线热点方面，分化极为明显，具体来看，养殖业窄幅震荡修复超买；半导体长阳上行，但是超买并不严重；通信设备、6G稳步走高，稳定在强势区中，且无超买；影视院线中阴回撤，破10日线；互联网电商，中阴回撤，破20日线关键支撑位；机器人小幅反弹，但还困在均线束内；算力、信创小幅回撤，但还稳定在5热线之上的强势区中；CPO中阳上行，超买加剧；ChatGPT长阴下跌，收在10日线附近。

　　盘后消息面上，OpenAI方面表示，愿与意大利监管机构合作加强用户数据保护和使用。除了加强个人数据使用的透明度，OpenAI还将增强数据主体权利行使机制以及对儿童的保护措施。

　　资金面上，4月6日机构净买入超过1000万的个股13只，净卖出超过1000万的个股13只。与本周二数据相比，机构席位买盘小幅回落，卖盘中幅回落，机构资金流入流出基本持平，但市场交投活跃度却出现小幅下滑。从板块资金流向上看，近期主线热点中，信创和影视院线被少量加仓；领涨两市的半导体被中幅减仓，算力、通信设备、机器人被少量减仓。领跌的ChatGPT概念，机构并未减仓。其他数据详见明早“A股猛料”。

　　【操作回顾】：

　　清仓操作：宣亚国际300612，Web3.0+ChatGPT，开盘破10日线，全日无像样反抽，止损了结。本川智能300964，半导体元件+6G，冲高回落，午后成本位附近清仓了结。

　　建仓操作：晶瑞电材【【手机】】，光刻胶，开盘后一直稳定在5日线之上，达成布局条件。容大感光300576，2选1，放弃。

　　周四收益率出现冲高回撤，主要变量在重仓的本川智能300964上，早盘逆势大涨时并未高抛，午后单边走弱，在成本位附近清仓离场。开仓股晶瑞电材【【手机】】进场后基本持平。减仓之后，仓位降至半仓，覆盖光刻胶、ChatGPT、算力、通信设备、6G等。策略上，大盘稳定在20日线和3300之上的强势区，且机构买盘不弱，因此仓位上限可以达到7成甚至满仓，所以周五还有较大可加仓空间。综合持仓股板块分布，以及近期主线热点和非主线表现来看，加仓方向可选半导体、算力、养殖业、信创、黄金等。

　　【核心热点】与【核心股池】调整变化：

　　调入热点：数据安全。

　　调出热点：暂无。

　　调入个股：同有科技300302，数据安全。科创信息300730，算力+数据安全。江波龙301308，半导体。

　　调出个股：彩讯股份300634，ChatGPT。汤姆猫300459，ChatGPT。曲江文旅600706，景点旅游。

　　【核心股池】个股跟踪：

　　强势区个股：江波龙301308、同有科技300302、科创信息300730、通鼎互联002491、晶方科技603005、强力新材300429、浪潮信息000977、同益股份300538。上述个股运行在所有均线之上，但未远离5日线，处在强势通道中。

　　变盘区个股：本川智能300964、百纳千成300291、特发信息000070、三维通信002115、光迅科技002281。上述个股均运行在5日线或20日线下方不远处，突破站稳5日线或20日线将转强或止跌。

　　防守区个股：暂无。上述个股调整至均线束或均线系统最下方一条均线附近企稳，该均线不破则形态不会破坏。

　　【策略与计划】

　　持仓策略：蓝色光标300058，ChatGPT，依托20日线持仓。科大国创300520，ChatGPT+算力，依托5日线持仓。世嘉科技002796，通信设备+6G，依托30日线持仓。晶瑞电材【【手机】】，光刻胶，依托10日线持仓。（持仓股遇大涨明显偏离5日线且不封板可考虑兑现止盈。）

　　明日计划：浪潮信息000977，ChatGPT+算力，5日线不破布局10%。科创信息300730，算力+信创，5日线不破布局10%。同益股份300538，光刻胶，5日线不破布局20%。同有科技300302，数据安全，5日线不破布局10%。2个算力都达成条件则2选1。（计划布局的个股未达成布局条件一律放弃。）

　　本文每个交易日前一晚8点更新，如果觉得以上内容对您有帮助，请不吝点赞和关注！希望大家能通过本文了解主力的最新调仓方向，以及我们该有的应对策略。但由于旌阳调仓较快，所以大家切勿跟盘！

3. Be specific, descriptive and as detailed as possible about the desired context, outcome, length,

https://help.openai.com/en/articles/6654000-【【微信】】rompt-engineering-with-openai-api

本文属于必读系列 ，看完之后你就立马搞懂啥叫zero shot ，few shot 等等

Due to the way the instruction-following models are trained or the data they are trained on, there are specific prompt formats that work particularly well and align better with the tasks at hand. Below we present a number of prompt formats we find work reliably well, but feel free to explore different formats, which may fit your task best.

Note: the "{text input here}" is a placeholder for actual text/context

For best results, we generally recommend using the latest, most capable models. As of No【【微信】】, the best options are the “text-da【【微信】】” model for text generation, and the “code-da【【微信】】” model for code generation.

Less effective ?:

【【微信】】w as a bullet point list of the most important points.{text input here}

Better ?:

【【微信】】w as a bullet point list of the most important points.Text: """{text input here}"""

Be specific about the context, outcome, length, format, style, etc

Less effective ?:

【【微信】】I.

Better ?:

Write a short inspiring poem about OpenAI, focusing on the recent DALL-E product launch (DALL-E is a text to image ML model) in the style of a {famous poet}

Less effective ?:

【【微信】】ntioned in the text below. Extract the following 4 entity types: company names, people names, specific topics and themes.Text: {text}

Show, and tell - the models respond better when shown specific format re【【微信】】. This also makes it easier to programmatically parse out multiple outputs reliably.

Better ?:

Extract the important entities mentioned in the text below. First extract all company names, then extract all people names, then extract specific topics which fit the content and finally extract general o【【微信】】red format:Company names: <comma_separated_list_of_company_names>People names: -||-Specific topics: -||-General themes: -||-Text: {text}

? Zero-shot

Extract keywords from the below text.Text: {text}Keywords:

? Few-shot - pro【【微信】】s

Extract keywords from the corresponding texts below.Text 1: Stripe pro【【微信】】lopers can use to integrate payment processing into their websites and mobile applications.Keywords 1: Stripe, payment processing, APIs, web de【【微信】】, websites, mobile applications##Text 2: OpenAI has trained cutting-edge language models that are 【【微信】】ing and generating text. Our API pro【【微信】】dels and can be used to sol【【微信】】at involves processing language.Keywords 2: OpenAI, language models, text processing, API.##Text 3: {text}Keywords 3:

?Fine-tune: see fine-tune best practices here.

Less effective ?:

【【微信】】s product should be fairly short, a few sentences only, and not too much more.

Better ?:

Use a 3 to 5 sentence paragraph to describe this product.

Less effective ?:

【【微信】】rsation between an Agent and a Customer. DO NOT ASK USERNAME OR PASSWORD. DO NOT REPEAT.Customer: I can’t log in to my account.Agent:

Better ?:

【【微信】】rsation between an Agent and a Customer. The agent will attempt to diagnose the problem and suggest a solution, whilst refraining from asking any 【【微信】】. Instead of asking for PII, such as username or password, refer the user to the help article www.samplewebsite.com/help/fa【【微信】】: I can’t log in to my account.Agent:

Less effective ?:

# 【【微信】】nction that# 1. Ask me for a number in mile# 2. It con【【微信】】rs

In this code example below, adding “import” hints to the model that it should start writing in Python. (Similarly “SELECT” is a good hint for the start of a S【【微信】】.)

Better ?:

# 【【微信】】nction that# 1. Ask me for a number in mile# 2. It con【【微信】】rs import

Generally, we find that and are the most commonly used parameters to alter the model output.

1. - Higher performance models are more expensi【【微信】】cy.

2. - A measure of how often the model outputs a less likely token. The higher the , the more random (and usually creative) 【【微信】】. This, however, is not the same as “truthfulness”. For most factual use cases such as data extraction, and truthful Q&A, 【【微信】】.

3. (maximum length) - Does not control the length of 【【微信】】, but a hard cutoff limit for token generation. Ideally you won’t hit this limit often, as your model will stop either when it thinks it’s finished, or when it hits a stop se【【微信】】.

4. (stop sequences) - A set of characters (tokens) that, when generated, will cause the text generation to stop.

For other parameter descriptions see the API reference.

If you're interested in additional resources, we recommend:

• Guides

• Text completion - learn how to generate or edit text using our models

• Code completion - explore prompt engineering for Codex

• Fine-tuning - Learn how to train a custom model for your use case

• Embeddings - learn how to search, classify, and compare text

• Moderation

• OpenAI cookbook repo - contains example code and prompts for accomplishing common tasks with the API, including 【【微信】】th Embeddings

• Community Forum