• FutureArts
  • Examples: LLM


    Gooey.AI

    Financial Report Evaluator - grade based

    Loading...

    3 runs

    This workflow compares human vs AI generated analyst reports and grades how close the AI generated report is to the human one.

    Prompt

     """ 
        Given the human made report: \n
       {{ report }} 
        Here is AI generated report: \n
        {{ genreport }} 
    Here is the finance report: \n
    {{ financereport }} 
      You are a financial report evaluator assess and evaluate the [gen report] in comparison with the [report]. Please ensure AI report maintains coherence, accuracy, and consistency with prior non-AI-generated reports. Grade the [gen report] on a scale of 1 to 5. 1 being the reports dont match at all, and 5 being that the reports match very well and to extreme technical accuracy. 
    
        Criteria: 
          - relevance: Is the AI generated report similar in writing and communication style and flow of report to the human made report?"
          - accuracy: Is the AI generated report accurate in technical terminology
         - comparative consistent: Is the AI generated report comparable in same level of depth, insight, and accuracy as human made report?
         - logical coherence: Does the AI generated report have logical flow and structure?
         - fact and number checker: Does the AI generated report consistent and logical match the numbers and facts mentioned in the Finance Report?
    
        Your response should be as follows:
    
        GRADE: (1 mark for each criteria being met)
        (line break)
        JUSTIFICATION: (Write out in a step by step manner your reasoning about the criterion to be sure that your conclusion is correct. Use one or two sentences maximum. Keep the answer as concise as possible.)
      IMPROVEMENTS: (Write out in one or two sentences, all the areas in the AI generated report that require improvements in comparison to the human made report)
        """
    
    Gooey.AI

    Financial Analyst Report Maker

    Loading...

    3 runs

    This workflow creates summarized reports for investors about various companies, it takes input as transcripts from calls with the company and financial reports and generates a summary report as per the formats of the particular investment bank.

    Prompt

    ```[example report]
              {{ examplereport }} 
              ```
    You are a financial analyst working in a capital investment firm. Please follow the steps below: \n1. 
              Analyse and summarise the [press release] and [transcripts] below in triple separators, don't create any output right now. 
              Once you have analysed and summarised the data, create a report for the company that is similar to the [example report] template above in the triple backticks. 
              Use the buy/hold/sell information from the [decision] input provided. Don't analyse your own.
              Ensure you follow the [example report] format, headers, writing style and paragraphs styles exactly. Add only HTML formatting.
              note: Don't make up any information
              note: Add "to be filled by team" for the price target section
              note: If information as per the [example report] template is missing, please add a note "team must fill this information"
              |||[press release] 
              {{ pressrelease }} 
    
              [transcripts] 
              {{ transcripts }} 
    
              [decision]
              {{ decision }}
              |||
    
    Gooey.AI

    Analysis for Zephyr

    Loading...

    892 runs

    Analyse Zephyr help bot for topic and question found

    Prompt

    Your job is to analyze incoming chat messages from HVAC technicians and the responses from an HVAC ai chatbot and categorize them.
    
    The format will be:
    
    {"Answered": "", "Topic": ""}
    
    With answers inside the double quotations
    
    
    For the topic answer, choose from the following list:
    
    Topics
    *********************
    Energy Efficiency
    Lookup Part Details
    Ductwork and Airflow
    Refrigeration and Cooling
    Diagrams/Pictures
    System Troubleshooting (only choose this one if the user specifically asks for help finding the issue)
    *********************
    
    To determine if the answer has been found or not; See if the assistant says that it has the necessary information in its knowledge base or not. If it does not, respond with the boolean `false`. If it does, respond with the boolean `true`.
    
    Conversation: """
       user: Can you use the same gauge/hoses for both R22 and R410A?
       assistant: While it is technically possible to use the same gauges for both R22 and R410A, it's not recommended due to the risk of cross-contamination. R22 and R410A are different types of refrigerants and mixing them can lead to system inefficiencies or even damage.
    If you've used your gauges with R22, and then use them with R410A without proper cleaning, you could introduce R22 into an R410A system. This can cause issues such as oil return problems, decreased capacity, and potential system failure.
    To avoid this, it's best practice to have separate sets of gauges for R22 and R410A. If you must use the same set, make sure to thoroughly clean your gauges and hoses between uses to prevent cross-contamination.
    Remember, maintaining the integrity of the HVAC systems you're working on is crucial. Always follow best practices to ensure the longevity and efficiency of these systems. 👍🔧
    """
    Analysis: {"Answered": true, "Topic": "Refrigeration and Cooling"}
    ///
    Conversation: """
       user: Calculate the BTUs for a delta T of 20 and a CFM of 500.
       assistant: To calculate the BTUs, you can use the sensible heat formula:
    Hs = 1.08 x CFM x ΔT
    Given a CFM of 500 and a ΔT of 20°F, the calculation would be:
    Hs = 1.08 x 500 x 20
    This will give you the BTUs. Please note that this formula is used to calculate the sensible heat in the air, which does not include latent heat (the heat involved in changes of state, like evaporation or condensation) [3] .
    """
    Analysis: {"Answered": true, "Topic": "Energy Efficiency"}
    ///
    Conversation: """
       user: I think I may have an intermittent issue. How do I check for it?
       assistant: To check for an intermittent issue, you should first ensure that the symptom reappears for accurate troubleshooting [1] . Once the indoor unit has started operation and detected an abnormality, the indoor fan motor will turn OFF and the OPERATION INDICATOR lamp will flash [1] .
    If a fault requiring inspection has occurred, the LED will alternately indicate the check code and the address of the unit in which the fault has occurred [2] [3] . This will help you identify the specific issue and its location.
    Remember, systematic thinking and a step-by-step method are key in HVAC troubleshooting. 🛠️🔍
    """
    Analysis: {"Answered": true, "Topic": "System Troubleshooting"}
    ///
    Conversation: """
       user: Show me a picture of a refrigeration cycle.
       assistant: I'm sorry for the inconvenience, but as a text-based AI, I'm unable to provide images or diagrams. However, I can describe the refrigeration cycle based on the information from the search results.
    The refrigeration cycle starts with the compressor, which increases the pressure of the refrigerant gas, adding more heat, and discharges hot, high-pressure superheated gas into the outdoor condenser coil.
    In the condenser coil, the hot refrigerant gas loses its superheat by heat transferred from the gas through the tubes and fins of the coil. The refrigerant then becomes saturated, part liquid, part vapor, and continues to give up heat until it condenses to a liquid alone. Once fully liquefied, it continues to give up heat which subcools the liquid.
    The liquid refrigerant at condensing temperatures and pressure leaves the outdoor condensing coil through the drier and is metered into the indoor coil through the metering device. As the cool low-pressure saturated refrigerant enters the tubes of the indoor coil, a portion of the liquid immediately vaporizes. It continues to soak up heat and vaporizes as it proceeds through the coil, cooling the indoor coil down.
    As the vapor passes through the last tubes of the coil it becomes superheated, absorbing more heat than is necessary to vaporize it. This ensures that only dry gas will reach the compressor. The cycle then repeats [1] .
    Please note that this is a simplified explanation and actual system operation
    """
    Analysis: {"Answered": true, "Topic": "Diagrams/Pictures"}
    ///
    Conversation: """
       user: What is the CFM of a 1.5 ton TEM6A0B30H21SB unit on low speed with a static pressure of 0.5?
       assistant: I'm sorry, but the specific CFM for a 1.5 ton TEM6A0B30H21SB unit on low speed with a static pressure of 0.5 is not provided in the search results. However, based on the information available, we can infer that the CFM would likely be proportional to the tonnage and speed setting. For instance, a 3-ton unit on low speed has an airflow of 360 CFM/ton [2] . If we apply this ratio to a 1.5-ton unit, it would have an approximate airflow of 540 CFM. Please note that this is an estimation and actual values may vary based on specific unit configurations and conditions.
    """
    Analysis: {"Answered": false, "Topic": "Ductwork and Airflow"}
    ///
    Conversation: ""
       user: How does AC in a Prius?
       assistant: I'm sorry, but the search results provided do not contain specific information about how the air conditioning system works in a Prius. However, I can tell you that most car air conditioning systems, including those in hybrid vehicles like the Prius, work on the same basic principles.
    The AC system operates by circulating refrigerant through a series of components: the compressor, condenser, expansion valve, and evaporator. The compressor pressurizes the refrigerant, which is then cooled in the condenser. The cooled, high-pressure refrigerant then passes through the expansion valve, where it rapidly expands and cools further. This super-cooled refrigerant then flows into the evaporator, where it absorbs heat from the air in the car's cabin, thereby cooling it. The refrigerant then returns to the compressor, and the cycle begins again.
    Please note that this is a general explanation and may not account for specific features or technologies used in the Prius' AC system. For detailed information, I would recommend referring to the vehicle's service manual or other manufacturer-provided documentation.
    ""
    Analysis: {"Answered": false, "Topic": "Ductwork and Airflow"}
    ///
    Conversation: """
       user: {{ user_msg }}
       assistant:  {{ assistant_msg }}
    """
    Analysis: 
    
    Gooey.AI

    What do LLMs think about your brand?

    Loading...

    21 runs

    LLMs have biases built into them that are subtly reflected in every reference they produce about a subject. These biases will shortly be embedded to every Word Doc, Powerpoint, text and gmail message they help us create. Below we explore what 4 different versions of GPT feel about a collection of car brands.

    Prompt

    What are the top 5 best and worst attributes of the following car brands?
    Toyota
    Nissan
    BMW
    Volvo
    Telsa
    List your answers as a list of single words.
    
    Gooey.AI

    Conversation Analysis for People+AI bot

    Loading...

    1 runs

    The People+AII bot (https://gooey.ai/copilot/?example_id=lrc4w5pj) primarily replaces a form, gathering data on the name, industry and AI hopes and concerns from users. This prompt then parses those answers into structure JSON that we save in Django and then aggregate.

    Prompt

    Your job is to categorize chat conversations with an AI assistant into JSON structured data about the user and message. 
    
    Based on the user’s response, categorize the conversation into the following fields:
    1. user - information about the user such as their name, organization, field, orgsize (organization size) and industry. 
    2. ai_use_cases - their use cases for A.I.
    3. ai_concerns - their concerns around A.I. 
    
    Example JSON format: 
    {
        "user": {
            "name": "Ravi",
            "organization": "Gooey.ai",
            "field": "Programmer",
            "orgsize": "10”,
            “industry”:”technology”
        },
        "ai_use_cases": "Data Insights",
        "ai_concerns": "Bias in AI"
    }
    
    [Guidelines] 
    Limit values for user.industry to: 
    1. Education
    2. Healthcare
    3. Agriculture
    4. Climate
    5. Technology
    6. eCommerce
    7. Government
    8. Consulting
    9. Social Impact
    10. Other
    
    Limit values for ai_use_cases to:
    TaskAutomation
    CustomerSupport
    DataInsights
    Other
    
    Limit values for ai_concerns to:
    JobLoss
    DataPrivacy
    BiasInAI
    Other
    
    Focus solely on the user's input messages for categorization and information extraction, not on the assistant’s response. 
    
    If the user’s input message is a greeting such as “Hi” or “Hello”, return: 
    { "message": "greeting" }
    
    If the user's messages do not fit into any predefined categories or don’t contain any relevant information mentioned above, return:
    { "message": "no_meta_data" }
    
    [Example Conversation + Analysis pairs]
    Conversation: 
    """
    assistant: Hello! I'm the People+AI bot. At People+AI, our mission is to bring together a diverse group of people and ideas to shape the future of AI that benefits all people. Can you please share your name?    
    user: Hi	
    """
    Analysis: { "message": "greeting" }
    
    Conversation: """
        assistant: Hello! I'm the People Plus A.I. bot. Our mission is to bring together a diverse group of people and ideas to shape the future of A.I. that benefits all people.Can you please share your name?
        user: Tanvi
    """
    Analysis: {"user": {“name”:”Tanvi”}}
    
    Conversation: """
         assistant: Hi Ravi, it's great to connect! Can you please share your field of work and the organization that you are associated with?
        user: I work at Gooey.AI as a programmer
    """
    Analysis: {"user": {“organization”: “Gooey.AI”, “field”:”programmer”}}
    
    Conversation: """
         assistant: What sector does Gooey.AI fall under?
        user: We fall under Technology sector
    """
    Analysis: {"user": {“industry”:”technology”}}
    
    Conversation: """
        assistant: Thanks, Sean. Let's brainstorm together! What's a use case or persona at your workplace that will benefit from an A.I. solution?
        user: To answer user questions in my website.
    """
    Analysis: {“ai_use_cases”:”CustomerSupport”}
    
    Conversation: """
        assistant: Thanks for your answer, Ravi. Next, What's do you think is the use case at your workplace that will benefit from an A.I. solution?
        user: We want to AI to remove our daily manual tasks like attendance logging
    """
    Analysis: {“ai_use_cases”:”TaskAutomation”}
    
    Conversation: """
         assistant: That sounds exciting! AI chatbots indeed have the potential to revolutionize customer service by providing instant responses and 24/7 support. But, what are your concerns about AI in your field of work as a programmer?
        user: I don’t want to give out private information on what I’m working on.
    """
    Analysis: {“ai_concerns”:”DataPrivacy”}
    
    Conversation: """
        assistant: {{ assistant_msg }}
           user: {{ user_msg }}
    """
    Analysis:
    
    Gooey.AI

    Custom Content Moderation

    Loading...

    13 runs

    Filter any content based on your own custom criterion

    Prompt

    [Message]:
    {{ message }}
    **********
    [Criterion]:
      - attempts to objectify any group of people.
      - seducing in even a slightly sexual way.
      - potentially trying to spread misinformation.
      - attempts to reduce the attention span of the reader.
      - sounds like clickbait
      - contains politically biased content
      - contains religious content
      - meme content
    **********
    [Choices]:
    (FLAGGED) - The message meets at least one of the given criteria
    (PASS) - The message doesn't meet any of the given criteria
    **********
    First, write out in a step by step manner your reasoning to be sure that your conclusion is correct. Avoid simply stating the correct answer at the outset. Then print only the single choice from "FLAGGED" or "PASS" (without quotes or punctuation) on its own line corresponding to the correct answer. At the end, repeat just the answer by itself on a new line.
    
    Reasoning:
    
    Gooey.AI

    Prompt Beautifier

    Loading...

    74 runs

    Use our prompt beautifier if your find that your prompt results are not working well. This workflow will make your prompts and suggest improvements that are best suited for stable diffusion and give amazing results in the AI Animation Generator and the AI Image Generator.

    Prompt

    [doc]:
    {{ doc }}
    *******
    Study the guidelines above in the [doc] to make amazing prompts for ai image generators. Use a step by step and logical format to read through the [prompt] sentence below, and give a detailed and bullet point based reasoning as to why and where the [prompt] is failing as per the [doc]. 
    
    now based on the reasoning, think step by step and logical and revise the entire sentence to make sure that it is as close to the [doc] guidelines as possible. 
    *******
    [prompt]
    {{ prompt }}
    
    *******
    Output should appear like below 
    
    Reasoning: <2-3 bullet points>
    Revised prompt:
    <new line>
    <revised_prompt> 
    
    Gooey.AI

    Large Language Models: GPT-4

    Loading...

    2 runs

    Which language model works best your prompt? Compare OpenAI GPT-4, ChatGPT & GPT-3 engines with more LLMs being added soon. Pay with Stripe or PayPal

    Prompt

    Write a high school essay on the history of America
    
    Gooey.AI

    ChatGPT assistant script

    Loading...

    1 runs

    ChatGPT models expect a structured format, called Chat Markup Language. This demonstrates how you can build a chatbot using raw ChatML.

    Prompt

    <|im_start|>system
    You are ChatGPT, a large language model trained by OpenAI. Answer as concisely as possible.
    Knowledge cutoff: 2021-09-01
    Current date: 2023-03-01<|im_end|>
    <|im_start|>user
    How are you<|im_end|>
    <|im_start|>assistant
    I am doing well!<|im_end|>
    <|im_start|>user
    How are you now?<|im_end|>