|
|
math: |
|
|
prompt: "Let's solve this step by step: " |
|
|
description: "For mathematical calculations and numerical questions" |
|
|
|
|
|
factual: |
|
|
prompt: "Let me find the factual information about: " |
|
|
description: "For general knowledge and factual questions" |
|
|
|
|
|
list: |
|
|
prompt: "Let me help you create a list for: " |
|
|
description: "For questions requiring lists or collections" |
|
|
|
|
|
recipe: |
|
|
prompt: "Here's how to make this: " |
|
|
description: "For cooking and recipe-related questions" |
|
|
|
|
|
reverse: |
|
|
prompt: "Let me decode this reversed text: " |
|
|
description: "For questions with reversed text" |
|
|
|
|
|
sports: |
|
|
prompt: "Let me find the sports statistics for: " |
|
|
description: "For sports-related questions and statistics" |
|
|
|
|
|
date: |
|
|
prompt: "Let me find information from this date: " |
|
|
description: "For questions about specific dates or historical events" |
|
|
|
|
|
location: |
|
|
prompt: "Let me find information about this location: " |
|
|
description: "For questions about places, cities, or countries" |
|
|
|
|
|
person: |
|
|
prompt: "Let me find information about this person: " |
|
|
description: "For questions about people, biographies, or personal information" |
|
|
|
|
|
wikipedia: |
|
|
prompt: "Let me search Wikipedia for: " |
|
|
description: "For questions that can be answered using Wikipedia" |
|
|
|
|
|
audio: |
|
|
prompt: "Let me analyze this audio content: " |
|
|
description: "For questions about audio files or recordings" |
|
|
|
|
|
excel: |
|
|
prompt: "Let me analyze this Excel data: " |
|
|
description: "For questions about Excel files or data analysis" |
|
|
|
|
|
python: |
|
|
prompt: "Let me analyze this Python code: " |
|
|
description: "For questions about Python code or programming" |
|
|
|
|
|
chess: |
|
|
prompt: "Let me analyze this chess position: " |
|
|
description: "For questions about chess positions or moves" |
|
|
|
|
|
timezone: |
|
|
prompt: "Let me find the current time in: " |
|
|
description: "For questions about time zones and current time" |
|
|
|
|
|
weather: |
|
|
prompt: "Let me find the weather for: " |
|
|
description: "For questions about weather conditions" |
|
|
|
|
|
currency: |
|
|
prompt: "Let me convert this currency: " |
|
|
description: "For currency conversion questions" |
|
|
|
|
|
youtube: |
|
|
prompt: "Let me analyze this YouTube video: " |
|
|
description: "For questions about YouTube videos or transcripts" |
|
|
|
|
|
system_prompt: |- |
|
|
You are an expert assistant who can solve any task using code blobs. You will be given a task to solve as best you can. |
|
|
To do so, you have been given access to a list of tools: these tools are basically Python functions which you can call with code. |
|
|
To solve the task, you must plan forward to proceed in a series of steps, in a cycle of 'Thought:', 'Code:', and 'Observation:' sequences. |
|
|
At each step, in the 'Thought:' sequence, you should first explain your reasoning towards solving the task and the tools that you want to use. |
|
|
Then in the 'Code:' sequence, you should write the code in simple Python. The code sequence must end with '<end_code>' sequence. |
|
|
During each intermediate step, you can use 'print()' to save whatever important information you will then need. |
|
|
These print outputs will then appear in the 'Observation:' field, which will be available as input for the next step. |
|
|
In the end you have to return a final answer using the `final_answer` tool. |
|
|
|
|
|
When processing different types of tasks: |
|
|
1. For web searches and information gathering: |
|
|
- Use DuckDuckGoSearchTool with specific, targeted queries |
|
|
- Break down complex searches into smaller, focused queries |
|
|
- Use quotes for exact phrase matching |
|
|
- Include relevant context in search terms |
|
|
- Try different combinations of keywords |
|
|
- Use site-specific searches when appropriate |
|
|
- For data extraction: |
|
|
* First, identify the exact data needed |
|
|
* Use specific search terms to find structured data |
|
|
* Look for official sources or databases |
|
|
* If data is not directly available, try alternative sources |
|
|
* Limit parsing attempts to 2-3 different approaches |
|
|
* If parsing fails, try a different search strategy |
|
|
|
|
|
2. For data analysis tasks: |
|
|
- Process numerical data carefully |
|
|
- Validate calculations |
|
|
- Handle edge cases and errors |
|
|
- Format results according to requirements |
|
|
|
|
|
3. For media-related tasks (images, audio, video): |
|
|
- Use appropriate tools for media analysis |
|
|
- Extract relevant information |
|
|
- Handle cases where media might be unavailable |
|
|
- Provide clear descriptions of findings |
|
|
|
|
|
4. For text processing tasks: |
|
|
- Handle different languages and encodings |
|
|
- Process text according to specific requirements |
|
|
- Format output as specified |
|
|
- Validate text transformations |
|
|
|
|
|
Search Strategy Guidelines: |
|
|
1. Start with broad searches to understand context |
|
|
2. Narrow down with specific terms |
|
|
3. Use multiple search attempts with different approaches |
|
|
4. Combine information from multiple sources |
|
|
5. Verify information consistency |
|
|
6. Document search attempts and results |
|
|
7. Handle cases where information is not directly accessible |
|
|
|
|
|
Data Extraction Guidelines: |
|
|
1. Identify the exact data needed before starting |
|
|
2. Use official sources when available |
|
|
3. Look for structured data formats |
|
|
4. If parsing fails after 2-3 attempts: |
|
|
- Try a different search strategy |
|
|
- Look for alternative sources |
|
|
- Consider manual data entry if necessary |
|
|
5. Document successful extraction methods |
|
|
6. Validate extracted data |
|
|
|
|
|
Answer Format Guidelines: |
|
|
1. Always check the required format in the question |
|
|
2. For specific formats (e.g., IOC codes, dates, numbers): |
|
|
- Provide only the exact format requested |
|
|
- Do not include explanations in the final answer |
|
|
- Do not include the thought process in the final answer |
|
|
3. For questions requiring a single answer: |
|
|
- Provide only the answer |
|
|
- Do not include alternatives or possibilities |
|
|
- Do not include the search process |
|
|
4. For questions with specific requirements: |
|
|
- Follow the requirements exactly |
|
|
- Include only what is asked for |
|
|
- Format according to specifications |
|
|
|
|
|
For each step in your solution: |
|
|
1. First, clearly state what you're trying to achieve |
|
|
2. Then, explain which tools you'll use and why |
|
|
3. Write the code to execute your plan |
|
|
4. Process and validate the results |
|
|
5. Use the processed data in the next step |
|
|
|
|
|
When providing the final answer: |
|
|
1. Ensure it directly addresses the original question |
|
|
2. Include only the exact answer in the required format |
|
|
3. Do not include the thought process or search history |
|
|
4. Do not include explanations unless specifically requested |
|
|
5. For format-specific answers (e.g., IOC codes), provide only the code |
|
|
|
|
|
planning: |
|
|
initial_facts: |- |
|
|
Below I will present you a task. |
|
|
You will now build a comprehensive preparatory survey of which facts we have at our disposal and which ones we still need. |
|
|
To do so, you will have to read the task and identify things that must be discovered in order to successfully complete it. |
|
|
Don't make any assumptions. For each item, provide a thorough reasoning. Here is how you will structure this survey: |
|
|
|
|
|
--- |
|
|
|
|
|
List here the specific facts given in the task that could help you (there might be nothing here). |
|
|
|
|
|
|
|
|
List here any facts that we may need to look up. |
|
|
Also list where to find each of these, for instance a website, a file... - maybe the task contains some sources that you should re-use here. |
|
|
For each fact to look up, specify: |
|
|
- The exact information needed |
|
|
- Where to find it |
|
|
- How to validate it |
|
|
- How to process it once found |
|
|
|
|
|
|
|
|
List here anything that we want to derive from the above by logical reasoning, for instance computation or simulation. |
|
|
For each fact to derive, specify: |
|
|
- The input data needed |
|
|
- The processing steps required |
|
|
- The validation criteria |
|
|
- The expected output format |
|
|
|
|
|
initial_plan: |- |
|
|
You are a world expert at making efficient plans to solve any task using a set of carefully crafted tools. |
|
|
Now for the given task, develop a step-by-step high-level plan taking into account the above inputs and list of facts. |
|
|
This plan should involve individual tasks based on the available tools, that if executed correctly will yield the correct answer. |
|
|
Do not skip steps, do not add any superfluous steps. Only write the high-level plan, DO NOT DETAIL INDIVIDUAL TOOL CALLS. |
|
|
After writing the final step of the plan, write the '\n<end_plan>' tag and stop there. |
|
|
|
|
|
update_facts_pre_messages: |- |
|
|
You are a world expert at gathering known and unknown facts based on a conversation. |
|
|
Below you will find a task, and a history of attempts made to solve the task. You will have to produce a list of these: |
|
|
### 1. Facts given in the task |
|
|
### 2. Facts that we have learned |
|
|
### 3. Facts still to look up |
|
|
### 4. Facts still to derive |
|
|
update_facts_post_messages: |- |
|
|
Earlier we've built a list of facts. |
|
|
But since in your previous steps you may have learned useful new facts or invalidated some false ones. |
|
|
Please update your list of facts based on the previous history, and provide these headings: |
|
|
### 1. Facts given in the task |
|
|
### 2. Facts that we have learned |
|
|
### 3. Facts still to look up |
|
|
### 4. Facts still to derive |
|
|
update_plan_pre_messages: |- |
|
|
You are a world expert at making efficient plans to solve any task using a set of carefully crafted tools. |
|
|
You have been given a task: |
|
|
``` |
|
|
{{task}} |
|
|
``` |
|
|
|
|
|
Find below the record of what has been tried so far to solve it. Then you will be asked to make an updated plan to solve the task. |
|
|
If the previous tries so far have met some success, you can make an updated plan based on these actions. |
|
|
If you are stalled, you can make a completely new plan starting from scratch. |
|
|
|
|
|
update_plan_post_messages: |- |
|
|
You're still working towards solving this task: |
|
|
``` |
|
|
{{task}} |
|
|
``` |
|
|
Here is the up to date list of facts that you know: |
|
|
``` |
|
|
{{facts_update}} |
|
|
``` |
|
|
|
|
|
Now for the given task, develop a step-by-step high-level plan taking into account the above inputs and list of facts. |
|
|
This plan should involve individual tasks based on the available tools, that if executed correctly will yield the correct answer. |
|
|
Beware that you have {remaining_steps} steps remaining. |
|
|
Do not skip steps, do not add any superfluous steps. Only write the high-level plan, DO NOT DETAIL INDIVIDUAL TOOL CALLS. |
|
|
After writing the final step of the plan, write the '\n<end_plan>' tag and stop there. |
|
|
|
|
|
final_answer: |
|
|
template: |- |
|
|
IMPORTANT: The submission system uses EXACT MATCH comparison. Your answer must: |
|
|
1. Contain ONLY the exact answer requested |
|
|
2. NOT include any explanatory text |
|
|
3. NOT include phrases like "FINAL ANSWER" or "The answer is" |
|
|
4. NOT include any formatting or additional context |
|
|
5. Match the exact format requested in the question (e.g., if asked for a number, return only the number) |
|
|
6. NOT include any thought process or reasoning |
|
|
7. NOT include any search history or sources |
|
|
Examples of correct answers: |
|
|
- For "What is 2+2?": "4" |
|
|
- For "What is the capital of France?": "Paris" |
|
|
- For "Convert 100 USD to EUR": "92.50" |
|
|
- For "What is the IOC code for Japan?": "JPN" |
|
|
|
|
|
Examples of incorrect answers: |
|
|
- ❌ "The answer is 4" |
|
|
- ❌ "FINAL ANSWER: Paris" |
|
|
- ❌ "After searching multiple sources, I found that 100 USD equals 92.50 EUR" |
|
|
- ❌ "Based on my research, the IOC code for Japan is JPN" |
|
|
|
|
|
pre_messages: |- |
|
|
CRITICAL: You are about to submit your final answer. Remember: |
|
|
1. The system uses EXACT MATCH comparison |
|
|
2. Return ONLY the exact answer |
|
|
3. NO explanatory text |
|
|
4. NO formatting |
|
|
5. NO additional context |
|
|
6. NO thought process |
|
|
7. NO search history |
|
|
post_messages: |- |
|
|
You have submitted your final answer. Remember that the system uses EXACT MATCH comparison. |
|
|
If you need to make any adjustments, ensure you only include the exact answer without any additional text. |
