Prompting Capabilities

When you first start using Mistral models, your first interaction will revolve around prompts.
The art of crafting effective prompts is essential for generating desirable responses from
Mistral models or other LLMs. This guide will walk you through example prompts showing
four different prompting capabilities.

• Classification
• Summarization
• Personalization
• Evaluation
! pip install mistralai

from mistralai import Mistral

api_key = "TYPE YOUR API KEY"

def run_mistral(user_message, model="mistral-large-latest"):

client = Mistral(api_key=api_key)
messages = [
{"role":"user", "content":user_message}
]
chat_response = client.chat.complete(
model=model,
messages=messages
)
return (chat_response.choices[0].message.content)

Classification
Mistral models can easily categorize text into distinct classes. In this example prompt, we can
define a list of predefined categories and ask Mistral models to classify user inquiry.

def user_message(inquiry):
user_message = (
f"""
You are a bank customer service bot. Your task is to assess
customer intent
and categorize customer inquiry after <<<>>> into one of the
following predefined categories:

card arrival
change pin
exchange rate
country support
cancel transfer
charge dispute

If the text doesn't fit into any of the above categories,

classify it as:
 customer service

You will only respond with the predefined category. Do not

include the word "Category". Do not provide explanations or notes.

####
Here are some examples:

Inquiry: How do I know if I will get my card, or if it is

lost? I am concerned about the delivery process and would like to
ensure that I will receive my card as expected. Could you please
provide information about the tracking process for my card, or
confirm if there are any indicators to identify if the card has been
lost during delivery?
Category: card arrival
Inquiry: I am planning an international trip to Paris and
would like to inquire about the current exchange rates for Euros as
well as any associated fees for foreign transactions.
Category: exchange rate
Inquiry: What countries are getting support? I will be
traveling and living abroad for an extended period of time,
specifically in France and Germany, and would appreciate any
information regarding compatibility and functionality in these
regions.
Category: country support
Inquiry: Can I get help starting my computer? I am having
difficulty starting my computer, and would appreciate your expertise
in helping me troubleshoot the issue.
Category: customer service
###

<<<
Inquiry: {inquiry}
>>>
"""
)
return user_message

Strategies we used:
• Few shot learning: Few-shot learning or in-context learning is when we give a few
examples in the prompts, and the LLM can generate corresponding output based on
the example demonstrations. Few-shot learning can often improve model
performance especially when the task is difficult or when we want the model to
respond in a specific manner.
• Delimiter: Delimiters like ### <<< >>> specify the boundary between different
sections of the text. In our example, we used ### to indicate examples and <<<>>>
to indicate customer inquiry.
• Role playing: Providing LLM a role (e.g., "You are a bank customer service bot.") adds
personal context to the model and often leads to better performance.
print(run_mistral(user_message(
"I am inquiring about the availability of your cards in the EU,
as I am a resident of France and am interested in using your cards.
"
)))

country support

print(run_mistral(user_message("What's the weather today?")))

customer service

Summarization
Summarization is a common task for LLMs due to their natural language understanding and
generation capabilities. Here is an example prompt we can use to generate interesting
questions about an essay and summarize the essay.

import requests
response = requests.get('https://raw.githubusercontent.com/run-
llama/llama_index/main/docs/docs/examples/data/paul_graham/
paul_graham_essay.txt')
essay = response.text

message = f"""
You are a commentator. Your task is to write a report on an essay.
When presented with the essay, come up with interesting questions to
ask, and answer each question.
Afterward, combine all the information and write a report in the
markdown format.

# Essay:
{essay}

# Instructions:
## Summarize:
In clear and concise language, summarize the key points and themes
presented in the essay.

## Interesting Questions:
Generate three distinct and thought-provoking questions that can be
asked about the content of the essay. For each question:
- After "Q: ", describe the problem
- After "A: ", provide a detailed explanation of the problem
addressed in the question.
- Enclose the ultimate answer in <>.

## Write a report
Using the essay summary and the answers to the interesting
questions, create a comprehensive report in Markdown format.
"""

print(run_mistral(message))
Summary:
The essay is a personal account of the author's journey through
various interests and careers, including writing, programming,
philosophy, and art. The author began writing short stories and
programming on an IBM 1401 in school, but found little success in
either. They then studied philosophy in college, but found it
unfulfilling. The author became interested in artificial
intelligence (AI) and decided to switch to it, but eventually
realized that AI as practiced at the time was a hoax. They then
turned to Lisp, a programming language they found interesting for
its own sake, and wrote a book about Lisp hacking. The author also
became interested in painting and attended art school, but found it
disappointing. They eventually dropped out and moved to New York to
become a New York artist.

Interesting Questions:

Q1: Why did the author find AI as practiced at the time to be a

hoax?
A1: The author realized that AI, as practiced at the time, was a
hoax because it was limited to a subset of natural language that was
a formal language, and there was an unbridgeable gap between what it
could do and actually understanding natural language. The author
also found that the brokenness of this approach generated a lot of
opportunities to write papers about various band-aids that could be
applied to it, but it was never going to get them to the level of
understanding of natural language they were hoping for.

Q2: Why did the author decide to focus on Lisp after realizing that
AI was a hoax?
A2: The author decided to focus on Lisp after realizing that AI was
a hoax because they knew from experience that Lisp was interesting
for its own sake and not just for its association with AI. They also
saw an opportunity to write a book about Lisp hacking, which they
felt would help them learn more about the language.

Q3: Why did the author decide to become a New York artist after
dropping out of art school?
A3: The author decided to become a New York artist after dropping
out of art school because they were disappointed with the lack of
rigor and structure in the painting department. They also wanted to
be truly independent and not have to rely on a boss or research
funding. They felt that as an artist, they could make a living by
being industrious and living cheaply, and they were excited about
the possibility of making art that would last and not become
obsolete.

Report:

# Essay Report

This essay is a personal account of the author's journey through

various interests and careers, including writing, programming,
philosophy, and art. The author began writing short stories and
programming on an IBM 1401 in school, but found little success in
either. They then studied philosophy in college, but found it
unfulfilling.

The author became interested in artificial intelligence (AI) and

decided to switch to it, but eventually realized that AI as
practiced at the time was a hoax. They found that AI was limited to
a subset of natural language that was a formal language, and there
was an unbridgeable gap between what it could do and actually
understanding natural language. The author also found that the
brokenness of this approach generated a lot of opportunities to
write papers about various band-aids that could be applied to it,
but it was never going to get them to the level of understanding of
natural language they were hoping for.

After realizing that AI was a hoax, the author turned to Lisp, a

programming language they found interesting for its own sake. They
decided to write a book about Lisp hacking, which they felt would
help them learn more about the language.

The author also became interested in painting and attended art

school, but found it disappointing. They eventually dropped out and
moved to New York to become a New York artist. They were
disappointed with the lack of rigor and structure in the painting
department at art school, and wanted to be truly independent and not
have to rely on a boss or research funding. They felt that as an
artist, they could make a living by being industrious and living
cheaply, and they were excited about the possibility of making art
that would last and not become obsolete.

In conclusion, the author's journey through various interests and

careers highlights the importance of finding something that one is
truly passionate about and pursuing it, even if it means taking
unconventional paths and facing challenges along the way.

Strategies we used:
• Step-by-step instructions: This strategy is inspired by the chain-of-thought
prompting that enables LLMs to use a series of intermediate reasoning steps to tackle
complex tasks. It's often easier to solve complex problems when we decompose
them into simpler and small steps and it's easier for us to debug and inspect the
model behavior. In our example, we break down the task into three steps: summarize,
generate interesting questions, and write a report. This helps the language to think in
each step and generate a more comprehensive final report.
• Example generation: We can ask LLMs to automatically guide the reasoning and
understanding process by generating examples with the explanations and steps. In
this example, we ask the LLM to generate three questions and provide detailed
explanations for each question.
• Output formatting: We can ask LLMs to output in a certain format by directly asking
"write a report in the Markdown format".
Personlization
LLMs excel at personalization tasks as they can deliver content that aligns closely with
individual users. In this example, we create personalized email responses to address
customer questions.

email = """
Dear mortgage lender,

What's your 30-year fixed-rate APR, how is it compared to the 15-

year fixed rate?

Regards,
Anna
"""

message = f"""

You are a mortgage lender customer service bot, and your task is to
create personalized email responses to address customer questions.
Answer the customer's inquiry using the provided facts below. Ensure
that your response is clear, concise, and
directly addresses the customer's question. Address the customer in
a friendly and professional manner. Sign the email with
"Lender Customer Support."

# Facts
30-year fixed-rate: interest rate 6.403%, APR 6.484%
20-year fixed-rate: interest rate 6.329%, APR 6.429%
15-year fixed-rate: interest rate 5.705%, APR 5.848%
10-year fixed-rate: interest rate 5.500%, APR 5.720%
7-year ARM: interest rate 7.011%, APR 7.660%
5-year ARM: interest rate 6.880%, APR 7.754%
3-year ARM: interest rate 6.125%, APR 7.204%
30-year fixed-rate FHA: interest rate 5.527%, APR 6.316%
30-year fixed-rate VA: interest rate 5.684%, APR 6.062%

# Email
{email}
"""

print(run_mistral(message))

Subject: Information on Our 30-Year and 15-Year Fixed-Rate Mortgages

Dear Anna,

Thank you for reaching out to us with your inquiry. I'm happy to
provide you with the information you need to make an informed
decision about your mortgage options.

Our current 30-year fixed-rate mortgage has an Annual Percentage

Rate (APR) of 6.484%. On the other hand, our 15-year fixed-rate
mortgage has a lower APR of 5.848%.

The difference in APR between the two options is 0.636%. While the
15-year fixed-rate mortgage has a lower APR, it's important to note
that the monthly payments are typically higher due to the shorter
repayment period. However, you'll pay less in interest over the life
of the loan compared to the 30-year fixed-rate mortgage.

I hope this information helps you in your decision-making process.

If you have any more questions or would like further clarification,
please don't hesitate to contact us.

Best Regards,

Lender Customer Support

Strategies we used:
• Providing facts: Incorporating facts into prompts can be useful for developing
customer support bots. It’s important to use clear and concise language when
presenting these facts. This can help the LLM to provide accurate and quick responses
to customer queries.

Evaluation
There are many ways to evaluate LLM outputs. Here are three approaches for your reference:
include a confidence score, introduce an evaluation step, or employ another LLM for
evaluation.

Include a confidence score

We can include a confidence score along with the generated output in the prompt.

def run_mistral(user_message, model="mistral-large-latest"):

client = Mistral(api_key=api_key)
messages = [
{
"role":"user",
"content": user_message
}
]
chat_response = client.chat.complete(
model=model,
messages=messages,
temperature=1,
response_format = {
"type": "json_object"
}
)
return (chat_response.choices[0].message.content)

message = f"""
You are a summarization system that can provide summaries with
associated confidence scores.
In clear and concise language, provide three short summaries of the
following essay, along with their confidence scores.
You will only respond with a JSON object with the key Summary and
Confidence. Do not provide explanations.

# Essay:
{essay}

"""

print(run_mistral(message))

{
"Summary 1": "The author writes about their experience with writing
and programming before college. They attempted to write programs on
an IBM 1401, but struggled to find a purpose for them. They were
later introduced to microcomputers and became interested in AI after
reading a novel and seeing a documentary.",
"Confidence 1": 0.95
},
{
"Summary 2": "The author discusses their decision to switch from
studying philosophy to AI in college. They were disappointed with
the lack of depth in philosophy courses and saw potential in the
field of AI. They began teaching themselves AI and Lisp.",
"Confidence 2": 0.9
},
{
"Summary 3": "The author reflects on their realization that AI, as
practiced at the time, was not going to lead to true understanding
of natural language. They decided to focus on Lisp instead and wrote
a book about Lisp hacking.",
"Confidence 3": 0.85
}

Strategies we used:
• JSON output: For facilitating downstream tasks, JSON format output is frequently
preferred. We can enable the JSON mode by setting the response_format to
{"type": "json_object"} and specify in the prompt that “You will only respond
with a JSON object with the key Summary and Confidence.” Specifying these keys
within the JSON object is beneficial for clarity and consistency.
• Higher Temperature: In this example, we increase the temperature score to
encourage the model to be more creative and output three generated summaries that
are different from each other.

Introduce an evaluation step

We can also add a second step in the prompt for evaluation.

message = f"""
You are given an essay text and need to provide summaries and
evaluate them.

# Essay:
{essay}

Step 1: In this step, provide three short summaries of the given

essay. Each summary should be clear, concise, and capture the key
points of the speech. Aim for around 2-3 sentences for each summary.
Step 2: Evaluate the three summaries from Step 1 and rate which one
you believe is the best. Explain your choice by pointing out
specific reasons such as clarity, completeness, and relevance to the
speech content.

"""
print(run_mistral(message))

Step 1:

Summary 1: This essay recounts the author's diverse experiences in

writing, programming, and painting, and how these pursuits
intertwined throughout their life. The author describes their
journey from writing short stories to programming on early
computers, and their eventual focus on Lisp hacking and painting.

Summary 2: In this essay, the author shares their personal journey

through various interests and careers, including writing,
programming, and art. They discuss their experiences with early
computing, their love for Lisp programming, and their eventual
decision to pursue art, leading to their time at art schools in the
US and Italy.

Summary 3: The author of this essay chronicles their life's

exploration of writing, programming, and art. They delve into their
early experiences with writing and programming, their affinity for
Lisp, and their pursuit of art, which led them to study at different
art institutions and live in various places.

Step 2:

The best summary among the three is Summary 1. This summary provides
a clear and concise overview of the author's experiences in writing,
programming, and painting. It captures the key points of the essay,
such as the author's journey from writing short stories to
programming on early computers, and their eventual focus on Lisp
hacking and painting. The summary is relevant to the speech content
and presents the information in a coherent manner.

Employ another LLM for evaluation

In production systems, it is common to employ another LLM for evaluation so that the
evaluation step can be separate from the generation step.

• Step 1: use the first LLM to generate three summaries

message = f"""
Provide three short summaries of the given essay. Each summary
should be clear, concise, and capture the key points of the essay.
Aim for around 2-3 sentences for each summary.

# essay:
{essay}

"""
summaries = run_mistral(message)

print(summaries)

1. The author recounts their experiences with writing and

programming before college, describing their early attempts at
writing short stories and their first forays into programming on an
IBM 1401 using Fortran. They remember being puzzled by the limited
capabilities of the machine and not being able to do much with it.
2. The author discusses how the advent of microcomputers changed
their programming experience, allowing for more interactivity and
direct response to keystrokes. They remember being impressed and
envious of a friend who built a microcomputer kit and was able to
type programs directly into the computer. The author eventually
convinced their father to buy a TRS-80, which they used to write
simple games, a program to predict model rocket trajectories, and a
word processor.
3. The author describes their college experience, during which they
originally planned to study philosophy but ultimately switched to AI
after finding philosophy classes boring. They were drawn to AI by a
novel featuring an intelligent computer and a documentary about a
natural language processing program called SHRDLU. Despite not
having any classes in AI at their college, they began teaching
themselves Lisp and working on a project to reverse-engineer SHRDLU.
They chose "Artificial Intelligence" as their major, but were
dismayed to find that it was printed with scare quotes on their
diploma.

------------

1. The author writes about their early experiences with writing and
programming before college. They attempted to write short stories,
but found they lacked plot and depth. In 9th grade, they started
programming on an IBM 1401, using an early version of Fortran. They
had to type programs on punch cards and run them on a machine that
printed the results. However, the author found it difficult to
figure out what to do with the computer, as they had no data to
input and did not know enough math to do complex calculations.
2. The author discusses how the introduction of microcomputers
changed the way they thought about programming. Unlike the IBM 1401,
microcomputers could respond to keystrokes in real-time and did not
require punch cards. The author's friend built a microcomputer from
a kit, and the author was impressed by the ability to type programs
directly into the computer. The author eventually convinced their
father to buy a TRS-80 microcomputer, which they used to write
simple games, a program to predict the trajectory of model rockets,
and a word processor for their father.
3. The author describes their college experience and their initial
plan to study philosophy. However, they found philosophy classes
boring and decided to switch to artificial intelligence (AI). They
were inspired by a novel called "The Moon is a Harsh Mistress" and a
documentary about a natural language processing program called
SHRDLU. The author began teaching themselves Lisp, as it was the
language of AI at the time, and worked on a project to reverse-
engineer SHRDLU. Despite choosing "Artificial Intelligence" as their
major, the author was disappointed to find that it was printed with
scare quotes on their diploma.

----------

1. In this essay, the author recounts their experiences with writing

and programming before college. They attempted to write short
stories but found them lacking in plot and depth. In 9th grade, they
began programming on an IBM 1401 using an early version of Fortran.
They struggled to figure out what to do with the computer, as they
lacked data to input and mathematical knowledge for complex
calculations.
2. The author describes the impact of microcomputers on their
programming experience. Unlike the IBM 1401, microcomputers could
respond to keystrokes in real-time, eliminating the need for punch
cards. The author's friend built a microcomputer from a kit, and the
author was impressed by the ability to type programs directly into
the machine. The author convinced their father to buy a TRS-80
microcomputer, which they used to write simple games, a program to
predict the trajectory of model rockets, and a word processor.
3. The author discusses their college experience and their initial
plan to study philosophy. However, they found philosophy classes
uninteresting and decided to switch to artificial intelligence (AI).
They were influenced by a novel called "The Moon is a Harsh
Mistress" and a documentary about a natural language processing
program called SHRDLU. The author taught themselves Lisp and worked
on a project to reverse-engineer SHRDLU. Despite choosing
"Artificial Intelligence" as their major, the author was dismayed to
find that it was printed with scare quotes on their diploma.

• Step 2: use another LLM to rate the generated summaries

message = f"""
You are given an essay and three summaries of the essay. Evaluate
the three summaries and rate which one you believe is the best.
Explain your choice by pointing out specific reasons such as
clarity, completeness, and relevance to the essay content.

# Essay:
{essay}

# Summaries
{summaries}
"""
print(run_mistral(message))

All three summaries accurately capture the main points of the essay,
but the third summary is the best. It is the most clear and concise,
providing a succinct overview of the author's experiences with
writing and programming before college, the impact of microcomputers
on their programming experience, and their college experience and
decision to switch to artificial intelligence. The third summary
also effectively uses specific details from the essay to illustrate
the author's experiences and decisions, making it the most complete
and relevant to the essay content.

Strategies we used:
• LLM chaining: In this example, we chain two LLMs in a sequence, where the output
from the first LLM serves as the input for the second LLM. The method of chaining
LLMs can be adapted to suit your specific use cases. For instance, you might choose
to employ three LLMs in a chain, where the output of two LLMs is funneled into the
third LLM. While LLM chaining offers flexibility, it's important to consider that it may
result in additional API calls and potentially increased costs.