Following up on my previous post on prompting rules, here are a few other tips gathered from the deeplearning.ai course; this time, we focus on how the model needs time to "elaborate" what it is doing.
This principle is based on the fact that the LLM of choice could sometime "rush" to the conclusion of a duty, leading to incorrect or poor results.
Sometimes the model makes statement that are plausible but not true: this is described as "having hallucinations" and it is due to the fact that it doesn't know the boundary of its knowledge very well; it has been trained on a gargantuan amount of data, but it might have not memorised it very well.
We need to be aware of hallucinations, and we need to put in place tactics to mitigate them and to use the model capabilities to double check on its own statements.
The key solution is to force the LLM to perform some steps of reasoning and checking them before getting to the final point; we instruct the model to spend more time and perform more computation on the task.
Similarly to humans, if the model is asked to perform a too complex task in a short amount of time (or with too few words describing what it has to do), it will "guess" an answer, and that will likely be incorrect.
I will explain this concept in the following paragraph, using the function get_completion in the code snippets; if you are curious to know what the function does, and to see the settings to start experimenting with chatGPT, check out check out this earlier post.
Specify the steps required to complete a task
text = f"""
In a charming village, siblings Jack and Jill set out on \
a quest to fetch water from a hilltop \
well. As they climbed, singing joyfully, misfortune \
struck—Jack tripped on a stone and tumbled \
down the hill, with Jill following suit. \
Though slightly battered, the pair returned home to \
comforting embraces. Despite the mishap, \
their adventurous spirits remained undimmed, and they \
continued exploring with delight.
"""Having a paragraph that we want to translate and extract information from, we might use the following prompt:
prompt_1 = f"""
Perform the following actions:
1 - Summarize the following text delimited by triple \
backticks with 1 sentence.
2 - Translate the summary into French.
3 - List each name in the French summary.
4 - Output a json object that contains the following \
keys: french_summary, num_names.
Separate your answers with line breaks.
Text:
```{text}```
"""
response = get_completion(prompt_1)
print("Completion for prompt 1:")
print(response)And we will get this result:
Completion for prompt 1:
Two siblings, Jack and Jill, go on a quest to fetch water from a well on a hilltop, but misfortune strikes and they both tumble down the hill, returning home slightly battered but with their adventurous spirits undimmed.
Deux frères et sœurs, Jack et Jill, partent en quête d'eau d'un puits sur une colline, mais un malheur frappe et ils tombent tous les deux de la colline, rentrant chez eux légèrement meurtris mais avec leurs esprits aventureux intacts.
Noms: Jack, Jill.
{
"french_summary": "Deux frères et sœurs, Jack et Jill, partent en quête d'eau d'un puits sur une colline, mais un malheur frappe et ils tombent tous les deux de la colline, rentrant chez eux légèrement meurtris mais avec leurs esprits aventureux intacts.",
"num_names": 2
}Asking for a specific output format
In this case, we use the same prompt as the previous example, but we add output instructions so that the result is easier to read, inspect and use:
prompt_2 = f"""
Your task is to perform the following actions:
1 - Summarize the following text delimited by
<> with 1 sentence.
2 - Translate the summary into French.
3 - List each name in the French summary.
4 - Output a json object that contains the
following keys: french_summary, num_names.
Use the following format:
Text: <text to summarize>
Summary: <summary>
Translation: <summary translation>
Names: <list of names in Italian summary>
Output JSON: <json with summary and num_names>
Text: <{text}>
"""
response = get_completion(prompt_2)
print("\nCompletion for prompt 2:")
print(response)Completion for prompt 2:
Summary: Jack and Jill go on a quest to fetch water, but misfortune strikes and they tumble down the hill, returning home slightly battered but with their adventurous spirits undimmed.
Translation: Jack et Jill partent en quête d'eau, mais un malheur frappe et ils tombent de la colline, rentrant chez eux légèrement meurtris mais avec leurs esprits aventureux intacts.
Names: Jack, Jill
Output JSON: {"french_summary": "Jack et Jill partent en quête d'eau, mais un malheur frappe et ils tombent de la colline, rentrant chez eux légèrement meurtris mais avec leurs esprits aventureux intacts.", "num_names": 2}Force the model to elaborate on its own solutions
This tactic forces the model to work out a solution by itself before getting to a conclusion, and to check it before moving to the next step.
Let's look at this example, where the model is giving a wrong answer:
prompt = f"""
Determine if the student's solution is correct or not.
Question:
I'm building a solar power installation and I need \
help working out the financials.
- Land costs $100 / square foot
- I can buy solar panels for $250 / square foot
- I negotiated a contract for maintenance that will cost \
me a flat $100k per year, and an additional $10 / square \
foot
What is the total cost for the first year of operations
as a function of the number of square feet.
Student's Solution:
Let x be the size of the installation in square feet.
Costs:
1. Land cost: 100x
2. Solar panel cost: 250x
3. Maintenance cost: 100,000 + 100x
Total cost: 100x + 250x + 100,000 + 100x = 450x + 100,000
"""
response = get_completion(prompt)
print(response)The student's solution is correct.The following prompt adds the instructions mentioned at the beginning of this section to the task addressed to the model:
prompt = f"""
Your task is to determine if the student's solution \
is correct or not.
To solve the problem do the following:
- First, work out your own solution to the problem.
- Then compare your solution to the student's solution \
and evaluate if the student's solution is correct or not.
Don't decide if the student's solution is correct until
you have done the problem yourself.
Use the following format:
Question:
```
question here
```
Student's solution:
```
student's solution here
```
Actual solution:
```
steps to work out the solution and your solution here
```
Is the student's solution the same as actual solution \
just calculated:
```
yes or no
```
Student grade:
```
correct or incorrect
```
Question:
```
I'm building a solar power installation and I need help \
working out the financials.
- Land costs $100 / square foot
- I can buy solar panels for $250 / square foot
- I negotiated a contract for maintenance that will cost \
me a flat $100k per year, and an additional $10 / square \
foot
What is the total cost for the first year of operations \
as a function of the number of square feet.
```
Student's solution:
```
Let x be the size of the installation in square feet.
Costs:
1. Land cost: 100x
2. Solar panel cost: 250x
3. Maintenance cost: 100,000 + 100x
Total cost: 100x + 250x + 100,000 + 100x = 450x + 100,000
```
Actual solution:
"""
response = get_completion(prompt)
print(response)This time, the model managed to give a correct answer:
Let x be the size of the installation in square feet.
Costs:
1. Land cost: 100x
2. Solar panel cost: 250x
3. Maintenance cost: 100,000 + 10x
Total cost: 100x + 250x + 100,000 + 10x = 360x + 100,000
Is the student's solution the same as actual solution just calculated:
No
Student grade:
IncorrectI hope this was useful! Come back for other posts on Machine Learning, AI and Tech!
