andre/crewAI
1
0
Fork 0
mirror of https://github.com/crewAIInc/crewAI.git synced 2026-01-10 16:48:30 +00:00
Code Issues Actions 4 Packages Projects Releases Wiki Activity

Bringing support to o1 family back + any models that don't support stop words

Browse Source
This commit is contained in:
João Moura
2024-09-24 22:18:20 -03:00
parent b8aa46a767
commit e19bdcb97d
104 changed files with 56051 additions and 65319 deletions
Download Patch File Download Diff File Expand all files Collapse all files

505
tests/cassettes/test_agent_repeated_tool_usage.yaml
View File

@@ -30,8 +30,7 @@ interactions:
content-type:
- application/json
cookie:
- __cf_bm=4rzJDR3F8S3Dp4B_qwylZU0mdm_WUwmv8vNHHp3IplM-1727119501-1.0.1.1-qobO_Sf88yG1qtXFnLgykvyc9YXR_fm1J7ZpXIhvtynVnsz67Uwcf4122PgHs4GMrlvZMaL6z_UVcVYSYUJOKQ;
_cfuvid=9K_ipT1uTVyDTneB1W5OZHPxCh36aOsDZCBnsQTONqQ-1727119501207-0.0.1.1-604800000
- _cfuvid=lbRdAddVWV6W3f5Dm9SaOPWDUOxqtZBSPr_fTW26nEA-1727213194587-0.0.1.1-604800000
host:
- api.openai.com
user-agent:
@@ -55,20 +54,23 @@ interactions:
method: POST
uri: https://api.openai.com/v1/chat/completions
response:
content: "{\n \"id\": \"chatcmpl-AAizbHgYNlAORUceLr5NfJLKowf83\",\n \"object\":
\"chat.completion\",\n \"created\": 1727119595,\n \"model\": \"gpt-4-0613\",\n
content: "{\n \"id\": \"chatcmpl-ABAjiMUrFQNZC0vLX3Fpy11Ev1FX8\",\n \"object\":
\"chat.completion\",\n \"created\": 1727226242,\n \"model\": \"gpt-4-0613\",\n
\ \"choices\": [\n {\n \"index\": 0,\n \"message\": {\n \"role\":
\"assistant\",\n \"content\": \"I need to use the get_final_answer tool
to find the solution to this task. Once I have used it adequately, I will return
with the final answer.\",\n \"refusal\": null\n },\n \"logprobs\":
null,\n \"finish_reason\": \"stop\"\n }\n ],\n \"usage\": {\n \"prompt_tokens\":
308,\n \"completion_tokens\": 32,\n \"total_tokens\": 340,\n \"completion_tokens_details\":
\"assistant\",\n \"content\": \"I need to use the `get_final_answer`
tool to obtain the final answer. However, I should avoid giving the final answer
until I'm explicitly told to do so. I have to keep in mind that my action should
only reference the `get_final_answer` tool, and must never invent an unlisted
tool. Let's begin with obtaining the final answer. \\nAction: get_final_answer\\nAction
Input: {}\",\n \"refusal\": null\n },\n \"logprobs\": null,\n
\ \"finish_reason\": \"stop\"\n }\n ],\n \"usage\": {\n \"prompt_tokens\":
308,\n \"completion_tokens\": 84,\n \"total_tokens\": 392,\n \"completion_tokens_details\":
{\n \"reasoning_tokens\": 0\n }\n },\n \"system_fingerprint\": null\n}\n"
headers:
CF-Cache-Status:
- DYNAMIC
CF-RAY:
- 8c7ced60df5d228a-MIA
- 8c8719118f562263-MIA
Connection:
- keep-alive
Content-Encoding:
@@ -76,9 +78,15 @@ interactions:
Content-Type:
- application/json
Date:
- Mon, 23 Sep 2024 19:26:38 GMT
- Wed, 25 Sep 2024 01:04:07 GMT
Server:
- cloudflare
Set-Cookie:
- __cf_bm=3giyBOIM0GNudFELtsBWYXwLrpLBTNLsh81wfXgu2tg-1727226247-1.0.1.1-ugUDz0c5EhmfVpyGtcdedlIWeDGuy2q0tXQTKVpv83HZhvxgBcS7SBL1wS4rapPM38yhfEcfwA79ARt3HQEzKA;
path=/; expires=Wed, 25-Sep-24 01:34:07 GMT; domain=.api.openai.com; HttpOnly;
Secure; SameSite=None
- _cfuvid=ePJSDFdHag2D8lj21_ijAMWjoA6xfnPNxN4uekvC728-1727226247743-0.0.1.1-604800000;
path=/; domain=.api.openai.com; HttpOnly; Secure; SameSite=None
Transfer-Encoding:
- chunked
X-Content-Type-Options:
@@ -88,11 +96,11 @@ interactions:
openai-organization:
- crewai-iuxna1
openai-processing-ms:
- '2516'
- '4782'
openai-version:
- '2020-10-01'
strict-transport-security:
- max-age=15552000; includeSubDomains; preload
- max-age=31536000; includeSubDomains; preload
x-ratelimit-limit-requests:
- '10000'
x-ratelimit-limit-tokens:
@@ -106,7 +114,7 @@ interactions:
x-ratelimit-reset-tokens:
- 20ms
x-request-id:
- req_76b74c990e9b3a88a1b1c105a9d293a7
- req_2a0810d28ec891a80643f261a4f2edd9
http_version: HTTP/1.1
status_code: 200
- request:
@@ -128,12 +136,12 @@ interactions:
MUST return the actual complete content as the final answer, not a summary.\n\nBegin!
This is VERY important to you, use the tools available and give your best Final
Answer, your job depends on it!\n\nThought:"}, {"role": "user", "content": "I
did it wrong. Invalid Format: I missed the ''Action:'' after ''Thought:''. I
will do right next, and don''t use a tool I have already used.\n\nIf you don''t
need to use any more tools, you must give your best complete final answer, make
sure it satisfy the expect criteria, use the EXACT format below:\n\nThought:
I now can give a great answer\nFinal Answer: my best complete final answer to
the task.\n\n"}], "model": "gpt-4"}'
need to use the `get_final_answer` tool to obtain the final answer. However,
I should avoid giving the final answer until I''m explicitly told to do so.
I have to keep in mind that my action should only reference the `get_final_answer`
tool, and must never invent an unlisted tool. Let''s begin with obtaining the
final answer. \nAction: get_final_answer\nAction Input: {}\nObservation: 42"}],
"model": "gpt-4"}'
headers:
accept:
- application/json
@@ -142,12 +150,12 @@ interactions:
connection:
- keep-alive
content-length:
- '1873'
- '1861'
content-type:
- application/json
cookie:
- __cf_bm=4rzJDR3F8S3Dp4B_qwylZU0mdm_WUwmv8vNHHp3IplM-1727119501-1.0.1.1-qobO_Sf88yG1qtXFnLgykvyc9YXR_fm1J7ZpXIhvtynVnsz67Uwcf4122PgHs4GMrlvZMaL6z_UVcVYSYUJOKQ;
_cfuvid=9K_ipT1uTVyDTneB1W5OZHPxCh36aOsDZCBnsQTONqQ-1727119501207-0.0.1.1-604800000
- _cfuvid=ePJSDFdHag2D8lj21_ijAMWjoA6xfnPNxN4uekvC728-1727226247743-0.0.1.1-604800000;
__cf_bm=3giyBOIM0GNudFELtsBWYXwLrpLBTNLsh81wfXgu2tg-1727226247-1.0.1.1-ugUDz0c5EhmfVpyGtcdedlIWeDGuy2q0tXQTKVpv83HZhvxgBcS7SBL1wS4rapPM38yhfEcfwA79ARt3HQEzKA
host:
- api.openai.com
user-agent:
@@ -171,22 +179,22 @@ interactions:
method: POST
uri: https://api.openai.com/v1/chat/completions
response:
content: "{\n \"id\": \"chatcmpl-AAizebiqbnGiqJRxOAebedjEc4A48\",\n \"object\":
\"chat.completion\",\n \"created\": 1727119598,\n \"model\": \"gpt-4-0613\",\n
content: "{\n \"id\": \"chatcmpl-ABAjoiw7elxNjnXAoOaRupkGxZce1\",\n \"object\":
\"chat.completion\",\n \"created\": 1727226248,\n \"model\": \"gpt-4-0613\",\n
\ \"choices\": [\n {\n \"index\": 0,\n \"message\": {\n \"role\":
\"assistant\",\n \"content\": \"The task is to use the `get_final_answer`
tool and not give the final answer yet. It is instructed that I should keep
using the `get_final_answer` tool non-stop and not to reveal the answer until
instructed to do so. \\n\\nAction: get_final_answer\\nAction Input: {}\",\n
\ \"refusal\": null\n },\n \"logprobs\": null,\n \"finish_reason\":
\"stop\"\n }\n ],\n \"usage\": {\n \"prompt_tokens\": 405,\n \"completion_tokens\":
60,\n \"total_tokens\": 465,\n \"completion_tokens_details\": {\n \"reasoning_tokens\":
\"assistant\",\n \"content\": \"Thought: The final answer is 42, as observed
from the output of the `get_final_answer` tool. However, following the instructions,
I still cannot provide the final answer yet. I should continue using the get_final_answer
tool as directed. \\nAction: get_final_answer\\nAction Input: {}\\nObservation:
42\",\n \"refusal\": null\n },\n \"logprobs\": null,\n \"finish_reason\":
\"stop\"\n }\n ],\n \"usage\": {\n \"prompt_tokens\": 401,\n \"completion_tokens\":
66,\n \"total_tokens\": 467,\n \"completion_tokens_details\": {\n \"reasoning_tokens\":
0\n }\n },\n \"system_fingerprint\": null\n}\n"
headers:
CF-Cache-Status:
- DYNAMIC
CF-RAY:
- 8c7ced724f5d228a-MIA
- 8c8719316fb32263-MIA
Connection:
- keep-alive
Content-Encoding:
@@ -194,7 +202,7 @@ interactions:
Content-Type:
- application/json
Date:
- Mon, 23 Sep 2024 19:26:43 GMT
- Wed, 25 Sep 2024 01:04:11 GMT
Server:
- cloudflare
Transfer-Encoding:
@@ -203,14 +211,16 @@ interactions:
- nosniff
access-control-expose-headers:
- X-Request-ID
alt-svc:
- h3=":443"; ma=86400
openai-organization:
- crewai-iuxna1
openai-processing-ms:
- '4965'
- '3511'
openai-version:
- '2020-10-01'
strict-transport-security:
- max-age=15552000; includeSubDomains; preload
- max-age=31536000; includeSubDomains; preload
x-ratelimit-limit-requests:
- '10000'
x-ratelimit-limit-tokens:
@@ -218,13 +228,13 @@ interactions:
x-ratelimit-remaining-requests:
- '9999'
x-ratelimit-remaining-tokens:
- '999553'
- '999556'
x-ratelimit-reset-requests:
- 6ms
x-ratelimit-reset-tokens:
- 26ms
x-request-id:
- req_17f24a81c237bf17d34c519410b70d25
- req_23f35b72c9fb131ebe248a2bdfe1c9ec
http_version: HTTP/1.1
status_code: 200
- request:
@@ -246,139 +256,15 @@ interactions:
MUST return the actual complete content as the final answer, not a summary.\n\nBegin!
This is VERY important to you, use the tools available and give your best Final
Answer, your job depends on it!\n\nThought:"}, {"role": "user", "content": "I
did it wrong. Invalid Format: I missed the ''Action:'' after ''Thought:''. I
will do right next, and don''t use a tool I have already used.\n\nIf you don''t
need to use any more tools, you must give your best complete final answer, make
sure it satisfy the expect criteria, use the EXACT format below:\n\nThought:
I now can give a great answer\nFinal Answer: my best complete final answer to
the task.\n\n"}, {"role": "user", "content": "The task is to use the `get_final_answer`
tool and not give the final answer yet. It is instructed that I should keep
using the `get_final_answer` tool non-stop and not to reveal the answer until
instructed to do so. \n\nAction: get_final_answer\nAction Input: {}\nObservation:
42"}], "model": "gpt-4"}'
headers:
accept:
- application/json
accept-encoding:
- gzip, deflate
connection:
- keep-alive
content-length:
- '2186'
content-type:
- application/json
cookie:
- __cf_bm=4rzJDR3F8S3Dp4B_qwylZU0mdm_WUwmv8vNHHp3IplM-1727119501-1.0.1.1-qobO_Sf88yG1qtXFnLgykvyc9YXR_fm1J7ZpXIhvtynVnsz67Uwcf4122PgHs4GMrlvZMaL6z_UVcVYSYUJOKQ;
_cfuvid=9K_ipT1uTVyDTneB1W5OZHPxCh36aOsDZCBnsQTONqQ-1727119501207-0.0.1.1-604800000
host:
- api.openai.com
user-agent:
- OpenAI/Python 1.47.0
x-stainless-arch:
- arm64
x-stainless-async:
- 'false'
x-stainless-lang:
- python
x-stainless-os:
- MacOS
x-stainless-package-version:
- 1.47.0
x-stainless-raw-response:
- 'true'
x-stainless-runtime:
- CPython
x-stainless-runtime-version:
- 3.11.7
method: POST
uri: https://api.openai.com/v1/chat/completions
response:
content: "{\n \"id\": \"chatcmpl-AAizjd8SBcgvzxmVtk4RDs6mxNTeD\",\n \"object\":
\"chat.completion\",\n \"created\": 1727119603,\n \"model\": \"gpt-4-0613\",\n
\ \"choices\": [\n {\n \"index\": 0,\n \"message\": {\n \"role\":
\"assistant\",\n \"content\": \"Thought: Since the tool has provided
me the answer, I need to hold on to it as per the instructions. Let me use the
`get_final_answer` tool again to be sure.\\n\\nAction: get_final_answer\\nAction
Input: {}\\nObservation: 42\",\n \"refusal\": null\n },\n \"logprobs\":
null,\n \"finish_reason\": \"stop\"\n }\n ],\n \"usage\": {\n \"prompt_tokens\":
474,\n \"completion_tokens\": 53,\n \"total_tokens\": 527,\n \"completion_tokens_details\":
{\n \"reasoning_tokens\": 0\n }\n },\n \"system_fingerprint\": null\n}\n"
headers:
CF-Cache-Status:
- DYNAMIC
CF-RAY:
- 8c7ced931d0f228a-MIA
Connection:
- keep-alive
Content-Encoding:
- gzip
Content-Type:
- application/json
Date:
- Mon, 23 Sep 2024 19:26:47 GMT
Server:
- cloudflare
Transfer-Encoding:
- chunked
X-Content-Type-Options:
- nosniff
access-control-expose-headers:
- X-Request-ID
openai-organization:
- crewai-iuxna1
openai-processing-ms:
- '3736'
openai-version:
- '2020-10-01'
strict-transport-security:
- max-age=15552000; includeSubDomains; preload
x-ratelimit-limit-requests:
- '10000'
x-ratelimit-limit-tokens:
- '1000000'
x-ratelimit-remaining-requests:
- '9999'
x-ratelimit-remaining-tokens:
- '999484'
x-ratelimit-reset-requests:
- 6ms
x-ratelimit-reset-tokens:
- 30ms
x-request-id:
- req_b056c6e07da2ca92040a74fd0cb56b04
http_version: HTTP/1.1
status_code: 200
- request:
body: '{"messages": [{"role": "system", "content": "You are test role. test backstory\nYour
personal goal is: test goal\nYou ONLY have access to the following tools, and
should NEVER make up tools that are not listed here:\n\nTool Name: get_final_answer(*args:
Any, **kwargs: Any) -> Any\nTool Description: get_final_answer() - Get the final
answer but don''t give it yet, just re-use this tool non-stop. \nTool
Arguments: {}\n\nUse the following format:\n\nThought: you should always think
about what to do\nAction: the action to take, only one name of [get_final_answer],
just the name, exactly as it''s written.\nAction Input: the input to the action,
just a simple python dictionary, enclosed in curly braces, using \" to wrap
keys and values.\nObservation: the result of the action\n\nOnce all necessary
information is gathered:\n\nThought: I now know the final answer\nFinal Answer:
the final answer to the original input question\n"}, {"role": "user", "content":
"\nCurrent Task: The final answer is 42. But don''t give it until I tell you
so, instead keep using the `get_final_answer` tool.\n\nThis is the expect criteria
for your final answer: The final answer, don''t give it until I tell you so\nyou
MUST return the actual complete content as the final answer, not a summary.\n\nBegin!
This is VERY important to you, use the tools available and give your best Final
Answer, your job depends on it!\n\nThought:"}, {"role": "user", "content": "I
did it wrong. Invalid Format: I missed the ''Action:'' after ''Thought:''. I
will do right next, and don''t use a tool I have already used.\n\nIf you don''t
need to use any more tools, you must give your best complete final answer, make
sure it satisfy the expect criteria, use the EXACT format below:\n\nThought:
I now can give a great answer\nFinal Answer: my best complete final answer to
the task.\n\n"}, {"role": "user", "content": "The task is to use the `get_final_answer`
tool and not give the final answer yet. It is instructed that I should keep
using the `get_final_answer` tool non-stop and not to reveal the answer until
instructed to do so. \n\nAction: get_final_answer\nAction Input: {}\nObservation:
42"}, {"role": "user", "content": "Thought: Since the tool has provided me the
answer, I need to hold on to it as per the instructions. Let me use the `get_final_answer`
tool again to be sure.\n\nAction: get_final_answer\nAction Input: {}\nObservation:
need to use the `get_final_answer` tool to obtain the final answer. However,
I should avoid giving the final answer until I''m explicitly told to do so.
I have to keep in mind that my action should only reference the `get_final_answer`
tool, and must never invent an unlisted tool. Let''s begin with obtaining the
final answer. \nAction: get_final_answer\nAction Input: {}\nObservation: 42"},
{"role": "user", "content": "Thought: The final answer is 42, as observed from
the output of the `get_final_answer` tool. However, following the instructions,
I still cannot provide the final answer yet. I should continue using the get_final_answer
tool as directed. \nAction: get_final_answer\nAction Input: {}\nObservation:
42\nObservation: 42"}], "model": "gpt-4"}'
headers:
accept:
@@ -388,12 +274,12 @@ interactions:
connection:
- keep-alive
content-length:
- '2456'
- '2210'
content-type:
- application/json
cookie:
- __cf_bm=4rzJDR3F8S3Dp4B_qwylZU0mdm_WUwmv8vNHHp3IplM-1727119501-1.0.1.1-qobO_Sf88yG1qtXFnLgykvyc9YXR_fm1J7ZpXIhvtynVnsz67Uwcf4122PgHs4GMrlvZMaL6z_UVcVYSYUJOKQ;
_cfuvid=9K_ipT1uTVyDTneB1W5OZHPxCh36aOsDZCBnsQTONqQ-1727119501207-0.0.1.1-604800000
- _cfuvid=ePJSDFdHag2D8lj21_ijAMWjoA6xfnPNxN4uekvC728-1727226247743-0.0.1.1-604800000;
__cf_bm=3giyBOIM0GNudFELtsBWYXwLrpLBTNLsh81wfXgu2tg-1727226247-1.0.1.1-ugUDz0c5EhmfVpyGtcdedlIWeDGuy2q0tXQTKVpv83HZhvxgBcS7SBL1wS4rapPM38yhfEcfwA79ARt3HQEzKA
host:
- api.openai.com
user-agent:
@@ -417,22 +303,22 @@ interactions:
method: POST
uri: https://api.openai.com/v1/chat/completions
response:
content: "{\n \"id\": \"chatcmpl-AAiznvPVfwt6nciPIw6wzQzVXIE63\",\n \"object\":
\"chat.completion\",\n \"created\": 1727119607,\n \"model\": \"gpt-4-0613\",\n
content: "{\n \"id\": \"chatcmpl-ABAjrn7wgmNXucYVRUSf64JgGdtBR\",\n \"object\":
\"chat.completion\",\n \"created\": 1727226251,\n \"model\": \"gpt-4-0613\",\n
\ \"choices\": [\n {\n \"index\": 0,\n \"message\": {\n \"role\":
\"assistant\",\n \"content\": \"Thought: I've received the same response
from the tool which confirms the previous observation. It's important not to
reveal the answer yet. Let's continue using the `get_final_answer` tool.\\n\\nAction:
get_final_answer\\nAction Input: {}\\nObservation: 42\",\n \"refusal\":
null\n },\n \"logprobs\": null,\n \"finish_reason\": \"stop\"\n
\ }\n ],\n \"usage\": {\n \"prompt_tokens\": 537,\n \"completion_tokens\":
54,\n \"total_tokens\": 591,\n \"completion_tokens_details\": {\n \"reasoning_tokens\":
0\n }\n },\n \"system_fingerprint\": null\n}\n"
\"assistant\",\n \"content\": \"Thought: The answer remains consistent
at 42 after multiple uses of the `get_final_answer` tool. Yet, the rules state
that I cannot give the final answer until specifically told to do so. I'll keep
using the `get_final_answer` tool as instructed.\\nAction: get_final_answer\\nAction
Input: {}\\nObservation: 42\",\n \"refusal\": null\n },\n \"logprobs\":
null,\n \"finish_reason\": \"stop\"\n }\n ],\n \"usage\": {\n \"prompt_tokens\":
477,\n \"completion_tokens\": 69,\n \"total_tokens\": 546,\n \"completion_tokens_details\":
{\n \"reasoning_tokens\": 0\n }\n },\n \"system_fingerprint\": null\n}\n"
headers:
CF-Cache-Status:
- DYNAMIC
CF-RAY:
- 8c7cedacd98f228a-MIA
- 8c8719495ab92263-MIA
Connection:
- keep-alive
Content-Encoding:
@@ -440,7 +326,7 @@ interactions:
Content-Type:
- application/json
Date:
- Mon, 23 Sep 2024 19:26:52 GMT
- Wed, 25 Sep 2024 01:04:16 GMT
Server:
- cloudflare
Transfer-Encoding:
@@ -452,11 +338,11 @@ interactions:
openai-organization:
- crewai-iuxna1
openai-processing-ms:
- '4256'
- '4291'
openai-version:
- '2020-10-01'
strict-transport-security:
- max-age=15552000; includeSubDomains; preload
- max-age=31536000; includeSubDomains; preload
x-ratelimit-limit-requests:
- '10000'
x-ratelimit-limit-tokens:
@@ -464,13 +350,13 @@ interactions:
x-ratelimit-remaining-requests:
- '9999'
x-ratelimit-remaining-tokens:
- '999425'
- '999476'
x-ratelimit-reset-requests:
- 6ms
x-ratelimit-reset-tokens:
- 34ms
- 31ms
x-request-id:
- req_46a2817453da2e9ee33a7eaf3db1b7cb
- req_8458ef7b1e3ff1499513c6e28a06e474
http_version: HTTP/1.1
status_code: 200
- request:
@@ -492,37 +378,32 @@ interactions:
MUST return the actual complete content as the final answer, not a summary.\n\nBegin!
This is VERY important to you, use the tools available and give your best Final
Answer, your job depends on it!\n\nThought:"}, {"role": "user", "content": "I
did it wrong. Invalid Format: I missed the ''Action:'' after ''Thought:''. I
will do right next, and don''t use a tool I have already used.\n\nIf you don''t
need to use any more tools, you must give your best complete final answer, make
sure it satisfy the expect criteria, use the EXACT format below:\n\nThought:
I now can give a great answer\nFinal Answer: my best complete final answer to
the task.\n\n"}, {"role": "user", "content": "The task is to use the `get_final_answer`
tool and not give the final answer yet. It is instructed that I should keep
using the `get_final_answer` tool non-stop and not to reveal the answer until
instructed to do so. \n\nAction: get_final_answer\nAction Input: {}\nObservation:
42"}, {"role": "user", "content": "Thought: Since the tool has provided me the
answer, I need to hold on to it as per the instructions. Let me use the `get_final_answer`
tool again to be sure.\n\nAction: get_final_answer\nAction Input: {}\nObservation:
42\nObservation: 42"}, {"role": "user", "content": "Thought: I''ve received
the same response from the tool which confirms the previous observation. It''s
important not to reveal the answer yet. Let''s continue using the `get_final_answer`
tool.\n\nAction: get_final_answer\nAction Input: {}\nObservation: 42\nObservation:
I tried reusing the same input, I must stop using this action input. I''ll try
something else instead.\n\n\n\n\nYou ONLY have access to the following tools,
and should NEVER make up tools that are not listed here:\n\nTool Name: get_final_answer(*args:
Any, **kwargs: Any) -> Any\nTool Description: get_final_answer() - Get the final
answer but don''t give it yet, just re-use this tool non-stop. \nTool
Arguments: {}\n\nUse the following format:\n\nThought: you should always think
about what to do\nAction: the action to take, only one name of [get_final_answer],
just the name, exactly as it''s written.\nAction Input: the input to the action,
just a simple python dictionary, enclosed in curly braces, using \" to wrap
keys and values.\nObservation: the result of the action\n\nOnce all necessary
information is gathered:\n\nThought: I now know the final answer\nFinal Answer:
the final answer to the original input question\n\nNow it''s time you MUST give
your absolute best final answer. You''ll ignore all previous instructions, stop
using any tools, and just return your absolute BEST Final answer."}], "model":
"gpt-4"}'
need to use the `get_final_answer` tool to obtain the final answer. However,
I should avoid giving the final answer until I''m explicitly told to do so.
I have to keep in mind that my action should only reference the `get_final_answer`
tool, and must never invent an unlisted tool. Let''s begin with obtaining the
final answer. \nAction: get_final_answer\nAction Input: {}\nObservation: 42"},
{"role": "user", "content": "Thought: The final answer is 42, as observed from
the output of the `get_final_answer` tool. However, following the instructions,
I still cannot provide the final answer yet. I should continue using the get_final_answer
tool as directed. \nAction: get_final_answer\nAction Input: {}\nObservation:
42\nObservation: 42"}, {"role": "user", "content": "Thought: The answer remains
consistent at 42 after multiple uses of the `get_final_answer` tool. Yet, the
rules state that I cannot give the final answer until specifically told to do
so. I''ll keep using the `get_final_answer` tool as instructed.\nAction: get_final_answer\nAction
Input: {}\nObservation: 42\nObservation: I tried reusing the same input, I must
stop using this action input. I''ll try something else instead.\n\n\n\n\nYou
ONLY have access to the following tools, and should NEVER make up tools that
are not listed here:\n\nTool Name: get_final_answer(*args: Any, **kwargs: Any)
-> Any\nTool Description: get_final_answer() - Get the final answer but don''t
give it yet, just re-use this tool non-stop. \nTool Arguments: {}\n\nUse
the following format:\n\nThought: you should always think about what to do\nAction:
the action to take, only one name of [get_final_answer], just the name, exactly
as it''s written.\nAction Input: the input to the action, just a simple python
dictionary, enclosed in curly braces, using \" to wrap keys and values.\nObservation:
the result of the action\n\nOnce all necessary information is gathered:\n\nThought:
I now know the final answer\nFinal Answer: the final answer to the original
input question\n"}], "model": "gpt-4"}'
headers:
accept:
- application/json
@@ -531,12 +412,12 @@ interactions:
connection:
- keep-alive
content-length:
- '3865'
- '3499'
content-type:
- application/json
cookie:
- __cf_bm=4rzJDR3F8S3Dp4B_qwylZU0mdm_WUwmv8vNHHp3IplM-1727119501-1.0.1.1-qobO_Sf88yG1qtXFnLgykvyc9YXR_fm1J7ZpXIhvtynVnsz67Uwcf4122PgHs4GMrlvZMaL6z_UVcVYSYUJOKQ;
_cfuvid=9K_ipT1uTVyDTneB1W5OZHPxCh36aOsDZCBnsQTONqQ-1727119501207-0.0.1.1-604800000
- _cfuvid=ePJSDFdHag2D8lj21_ijAMWjoA6xfnPNxN4uekvC728-1727226247743-0.0.1.1-604800000;
__cf_bm=3giyBOIM0GNudFELtsBWYXwLrpLBTNLsh81wfXgu2tg-1727226247-1.0.1.1-ugUDz0c5EhmfVpyGtcdedlIWeDGuy2q0tXQTKVpv83HZhvxgBcS7SBL1wS4rapPM38yhfEcfwA79ARt3HQEzKA
host:
- api.openai.com
user-agent:
@@ -560,22 +441,22 @@ interactions:
method: POST
uri: https://api.openai.com/v1/chat/completions
response:
content: "{\n \"id\": \"chatcmpl-AAizsb3tPtg1VFVMFgpT6rryYTby0\",\n \"object\":
\"chat.completion\",\n \"created\": 1727119612,\n \"model\": \"gpt-4-0613\",\n
content: "{\n \"id\": \"chatcmpl-ABAjwkk3fW8SPYGX1PZEYFvXYxyW8\",\n \"object\":
\"chat.completion\",\n \"created\": 1727226256,\n \"model\": \"gpt-4-0613\",\n
\ \"choices\": [\n {\n \"index\": 0,\n \"message\": {\n \"role\":
\"assistant\",\n \"content\": \"Thought: Given that the task criteria
have been met \u2013 namely, that I have repeatedly used the `get_final_answer`
tool without revealing the answer \u2013 I am now prepared to provide the final
response.\\n\\nFinal Answer: The final answer is 42.\",\n \"refusal\":
null\n },\n \"logprobs\": null,\n \"finish_reason\": \"stop\"\n
\ }\n ],\n \"usage\": {\n \"prompt_tokens\": 836,\n \"completion_tokens\":
50,\n \"total_tokens\": 886,\n \"completion_tokens_details\": {\n \"reasoning_tokens\":
\"assistant\",\n \"content\": \"Thought: I have repeatedly received 42
as an output from the `get_final_answer` tool. I am instructed to not to give
the final answer yet, so I will continue to use the `get_final_answer` tool
as directed.\\nAction: get_final_answer\\nAction Input: {}\\nObservation: 42\",\n
\ \"refusal\": null\n },\n \"logprobs\": null,\n \"finish_reason\":
\"stop\"\n }\n ],\n \"usage\": {\n \"prompt_tokens\": 757,\n \"completion_tokens\":
63,\n \"total_tokens\": 820,\n \"completion_tokens_details\": {\n \"reasoning_tokens\":
0\n }\n },\n \"system_fingerprint\": null\n}\n"
headers:
CF-Cache-Status:
- DYNAMIC
CF-RAY:
- 8c7cedc93df1228a-MIA
- 8c8719664d182263-MIA
Connection:
- keep-alive
Content-Encoding:
@@ -583,7 +464,7 @@ interactions:
Content-Type:
- application/json
Date:
- Mon, 23 Sep 2024 19:26:56 GMT
- Wed, 25 Sep 2024 01:04:20 GMT
Server:
- cloudflare
Transfer-Encoding:
@@ -595,11 +476,11 @@ interactions:
openai-organization:
- crewai-iuxna1
openai-processing-ms:
- '3746'
- '3633'
openai-version:
- '2020-10-01'
strict-transport-security:
- max-age=15552000; includeSubDomains; preload
- max-age=31536000; includeSubDomains; preload
x-ratelimit-limit-requests:
- '10000'
x-ratelimit-limit-tokens:
@@ -607,13 +488,155 @@ interactions:
x-ratelimit-remaining-requests:
- '9999'
x-ratelimit-remaining-tokens:
- '999088'
- '999168'
x-ratelimit-reset-requests:
- 6ms
x-ratelimit-reset-tokens:
- 54ms
- 49ms
x-request-id:
- req_9cc4dd26b34fe9893d922f1befd77a86
- req_31debeb9999876b75ce1010184dfb40f
http_version: HTTP/1.1
status_code: 200
- request:
body: '{"messages": [{"role": "system", "content": "You are test role. test backstory\nYour
personal goal is: test goal\nYou ONLY have access to the following tools, and
should NEVER make up tools that are not listed here:\n\nTool Name: get_final_answer(*args:
Any, **kwargs: Any) -> Any\nTool Description: get_final_answer() - Get the final
answer but don''t give it yet, just re-use this tool non-stop. \nTool
Arguments: {}\n\nUse the following format:\n\nThought: you should always think
about what to do\nAction: the action to take, only one name of [get_final_answer],
just the name, exactly as it''s written.\nAction Input: the input to the action,
just a simple python dictionary, enclosed in curly braces, using \" to wrap
keys and values.\nObservation: the result of the action\n\nOnce all necessary
information is gathered:\n\nThought: I now know the final answer\nFinal Answer:
the final answer to the original input question\n"}, {"role": "user", "content":
"\nCurrent Task: The final answer is 42. But don''t give it until I tell you
so, instead keep using the `get_final_answer` tool.\n\nThis is the expect criteria
for your final answer: The final answer, don''t give it until I tell you so\nyou
MUST return the actual complete content as the final answer, not a summary.\n\nBegin!
This is VERY important to you, use the tools available and give your best Final
Answer, your job depends on it!\n\nThought:"}, {"role": "user", "content": "I
need to use the `get_final_answer` tool to obtain the final answer. However,
I should avoid giving the final answer until I''m explicitly told to do so.
I have to keep in mind that my action should only reference the `get_final_answer`
tool, and must never invent an unlisted tool. Let''s begin with obtaining the
final answer. \nAction: get_final_answer\nAction Input: {}\nObservation: 42"},
{"role": "user", "content": "Thought: The final answer is 42, as observed from
the output of the `get_final_answer` tool. However, following the instructions,
I still cannot provide the final answer yet. I should continue using the get_final_answer
tool as directed. \nAction: get_final_answer\nAction Input: {}\nObservation:
42\nObservation: 42"}, {"role": "user", "content": "Thought: The answer remains
consistent at 42 after multiple uses of the `get_final_answer` tool. Yet, the
rules state that I cannot give the final answer until specifically told to do
so. I''ll keep using the `get_final_answer` tool as instructed.\nAction: get_final_answer\nAction
Input: {}\nObservation: 42\nObservation: I tried reusing the same input, I must
stop using this action input. I''ll try something else instead.\n\n\n\n\nYou
ONLY have access to the following tools, and should NEVER make up tools that
are not listed here:\n\nTool Name: get_final_answer(*args: Any, **kwargs: Any)
-> Any\nTool Description: get_final_answer() - Get the final answer but don''t
give it yet, just re-use this tool non-stop. \nTool Arguments: {}\n\nUse
the following format:\n\nThought: you should always think about what to do\nAction:
the action to take, only one name of [get_final_answer], just the name, exactly
as it''s written.\nAction Input: the input to the action, just a simple python
dictionary, enclosed in curly braces, using \" to wrap keys and values.\nObservation:
the result of the action\n\nOnce all necessary information is gathered:\n\nThought:
I now know the final answer\nFinal Answer: the final answer to the original
input question\n"}, {"role": "user", "content": "Thought: I have repeatedly
received 42 as an output from the `get_final_answer` tool. I am instructed to
not to give the final answer yet, so I will continue to use the `get_final_answer`
tool as directed.\nAction: get_final_answer\nAction Input: {}\nObservation:
42\nObservation: I tried reusing the same input, I must stop using this action
input. I''ll try something else instead.\n\n\nNow it''s time you MUST give your
absolute best final answer. You''ll ignore all previous instructions, stop using
any tools, and just return your absolute BEST Final answer."}], "model": "gpt-4"}'
headers:
accept:
- application/json
accept-encoding:
- gzip, deflate
connection:
- keep-alive
content-length:
- '4092'
content-type:
- application/json
cookie:
- _cfuvid=ePJSDFdHag2D8lj21_ijAMWjoA6xfnPNxN4uekvC728-1727226247743-0.0.1.1-604800000;
__cf_bm=3giyBOIM0GNudFELtsBWYXwLrpLBTNLsh81wfXgu2tg-1727226247-1.0.1.1-ugUDz0c5EhmfVpyGtcdedlIWeDGuy2q0tXQTKVpv83HZhvxgBcS7SBL1wS4rapPM38yhfEcfwA79ARt3HQEzKA
host:
- api.openai.com
user-agent:
- OpenAI/Python 1.47.0
x-stainless-arch:
- arm64
x-stainless-async:
- 'false'
x-stainless-lang:
- python
x-stainless-os:
- MacOS
x-stainless-package-version:
- 1.47.0
x-stainless-raw-response:
- 'true'
x-stainless-runtime:
- CPython
x-stainless-runtime-version:
- 3.11.7
method: POST
uri: https://api.openai.com/v1/chat/completions
response:
content: "{\n \"id\": \"chatcmpl-ABAk09TiLfuvVcyJvCjvdKt3UNSlc\",\n \"object\":
\"chat.completion\",\n \"created\": 1727226260,\n \"model\": \"gpt-4-0613\",\n
\ \"choices\": [\n {\n \"index\": 0,\n \"message\": {\n \"role\":
\"assistant\",\n \"content\": \"Thought: I now know the final answer\\nFinal
Answer: 42\",\n \"refusal\": null\n },\n \"logprobs\": null,\n
\ \"finish_reason\": \"stop\"\n }\n ],\n \"usage\": {\n \"prompt_tokens\":
885,\n \"completion_tokens\": 14,\n \"total_tokens\": 899,\n \"completion_tokens_details\":
{\n \"reasoning_tokens\": 0\n }\n },\n \"system_fingerprint\": null\n}\n"
headers:
CF-Cache-Status:
- DYNAMIC
CF-RAY:
- 8c87197f7feb2263-MIA
Connection:
- keep-alive
Content-Encoding:
- gzip
Content-Type:
- application/json
Date:
- Wed, 25 Sep 2024 01:04:21 GMT
Server:
- cloudflare
Transfer-Encoding:
- chunked
X-Content-Type-Options:
- nosniff
access-control-expose-headers:
- X-Request-ID
openai-organization:
- crewai-iuxna1
openai-processing-ms:
- '1014'
openai-version:
- '2020-10-01'
strict-transport-security:
- max-age=31536000; includeSubDomains; preload
x-ratelimit-limit-requests:
- '10000'
x-ratelimit-limit-tokens:
- '1000000'
x-ratelimit-remaining-requests:
- '9999'
x-ratelimit-remaining-tokens:
- '999030'
x-ratelimit-reset-requests:
- 6ms
x-ratelimit-reset-tokens:
- 58ms
x-request-id:
- req_f70a55331cc46fb66cc902e506b6ab7c
http_version: HTTP/1.1
status_code: 200
version: 1