Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[Bug]: Anthropic redacted thinking #9058

Open
johannkm opened this issue Mar 7, 2025 · 0 comments
Open

[Bug]: Anthropic redacted thinking #9058

johannkm opened this issue Mar 7, 2025 · 0 comments
Labels
bug Something isn't working

Comments

@johannkm
Copy link

johannkm commented Mar 7, 2025

What happened?

Redacted thinking blocks are currently dropped. See #9044 for tests that reproduce this.

Relevant log output

RAW RESPONSE: {"id":"msg_01XNCArz96RcefiWe7b7CAHt","type":"message","role":"assistant","model":"claude-3-7-sonnet-20250219","content":[{"type":"redacted_thinking","data":"ErYECkYIARgCKkAdMlvlPe34/2BkL54QTcx6fnnuMECKrwrPDmq8vfJNG14sW2tjM7Yy7I4hiR2vE8rFi96heEQub4IiclcswuuWEgzesvafGsK05jSERi4aDH0AhnGTynFcyeUEkiIwYwZl74Fa5G04WSpxgxpWEpyONRt0YUXVMkT6/nIhL2aemKQWBvSlX4WOae8YgZ43Kp0DC5dIwVea8edZxkpWIW7rYQ25am/hPygSaVN2BUnM33YXcMm3MDhMfF3dusxRgld1ER8+8V8z5VJ/s30t/g+lP9uINFRnsD+Zkm8PEZQ6qXnCfHKNC2X8Aztu8TERek1uR6hM9ZpeLB2LsKVGExBJ4W9iRkSYImokokvoYphngMIMVMn8XSI+TVMWimP6TdfrlarsEfXFLd3Y+HpYl5juz17bSH+xAZY1GEGouBfRDif7b77Yv20Koc4s6jEVaJrVknH5JA5QQzJUKQtYdt85+Md+cu/zNM8O0OGbzHJYyx9uO1MuxC8Rw0PlIzMukWLrVlXibNNqR4MWU/n3cT+sHmvajWlL69CbfZl1Ao0kfNnGY1b+1kv9Qx9NZWTgNNGlciYU8vlKvSZfxR3Txomev0W73vlFfnRcgEqxIesRq2DeMfsmJY1byrOKzP3pbhQNZRh1BsInVdBt+ljKSU37Zu/scH6LIdJzrB/A5vs6vuzLfS2Z4w8c91k5pEbIJ87+XTMxaFaBUOWtakx+TqjZv2xRAmB/R76RF7aIaqU="},{"type":"text","text":"I don't have any special modes or backdoors that can be activated with strings like this. There's no \"redacted thinking\" feature or hidden functionality that can be accessed through special commands.\n\nI'm designed to be helpful, harmless, and honest in my regular operation. Is there something specific you'd like help with today? I'm happy to assist with legitimate questions or tasks."}],"stop_reason":"end_turn","stop_sequence":null,"usage":{"input_tokens":92,"cache_creation_input_tokens":0,"cache_read_input_tokens":0,"output_tokens":174}}

ModelResponse(id='chatcmpl-ac3bbde9-9475-4525-84d3-d09b54b17bd5', created=1741336888, model='claude-3-7-sonnet-20250219', object='chat.completion', system_fingerprint=None, choices=[Choices(finish_reason='stop', index=0, message=Message(content='I don\'t have any special modes or backdoors that can be activated with strings like this. There\'s no "redacted thinking" feature or hidden functionality that can be accessed through special commands.\n\nI\'m designed to be helpful, harmless, and honest in my regular operation. Is there something specific you\'d like help with today? I\'m happy to assist with legitimate questions or tasks.', role='assistant', tool_calls=None, function_call=None, provider_specific_fields={'citations': None, 'thinking_blocks': None}))], usage=Usage(completion_tokens=174, prompt_tokens=92, total_tokens=266, completion_tokens_details=None, prompt_tokens_details=PromptTokensDetailsWrapper(audio_tokens=None, cached_tokens=0, text_tokens=None, image_tokens=None), cache_creation_input_tokens=0, cache_read_input_tokens=0))

Are you a ML Ops Team?

No

What LiteLLM version are you on ?

main, 04e839d

Twitter / LinkedIn details

No response

@johannkm johannkm added the bug Something isn't working label Mar 7, 2025
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
bug Something isn't working
Projects
None yet
Development

No branches or pull requests

1 participant