`claude-researcher` is a constrained agent — meaning its behavior is highly-controlled, leading to better results than open-ended agents. It chains together lots of Claude 3 calls (and Google access) that work together to create a detailed report on a topic of your choice.
PROMPT ENGINEERING
-
GPT-4 Prompts Reshape Academic Recommendation Letters
By
–
The future happened when a student asking for a letter of recommendation sent me the suggested GPT-4 prompt he wanted me to use to generate the letter itself. Likely where recommendation letters are eventually going to go: 1-10 scale & a single sentence assessment. No ceremony.
-
Programming LLMs with DSPy Instead of Prompting
By
–
Program your LLMs not prompting. Nice thread Akshay on DSPy!
-
Claude 3 Opus and Sonnet 200k Models Now Available
By
–
We've added Claude 3 Opus 200k and Sonnet 200k as bases for your prompt bots! These multimodal models are the most powerful base bots available, and can handle longer inputs for more natural back-and-forth conversations, in-depth analysis, multi-step tasks, and more. pic.twitter.com/QKjGlkCTjY
— Poe (@poe_platform) 5 avril 2024We've added Claude 3 Opus 200k and Sonnet 200k as bases for your prompt bots! These multimodal models are the most powerful base bots available, and can handle longer inputs for more natural back-and-forth conversations, in-depth analysis, multi-step tasks, and more.
-
Improve Foundation Model Performance Without Fine-Tuning
By
–
Learn how to get better foundation model performance without additional fine-tuning. This event is TODAY at Noon Pacific. Time is running out! https://
buff.ly/495ZtmN -
Refining LLM Outputs by Filtering Overused Marketing Language
By
–
Here is my filter list which is similar for all LLMs – they are all way too “exciting” 'exciting', 'enhanced', 'interaction', 'engage', 'simplifies', 'enhances', 'dynamic', 'interactions', 'easily', 'more effectively', 'effectively', 'streamlines', 'better user experience',
-
Mastering LLMs Requires Deep Learning and Intuitive Techniques
By
–
The greatest misconception about LLMs is the idea that they're easy to use Learning to use them effectively takes a great deal of effort, and is extremely unintuitive I've found it worth the investment personally, but wow there's a lot of depth to it!
-
System Prompt Reliability Against Prompt Injection Attacks
By
–
The ability to follow instructions in the system prompt and 100% reliably ignore conflicting instructions in the regular prompt, with hard proof that it works (Aka a fix for prompt injection)
-
Claude’s Million Token Context Enables 250+ Tool Support
By
–
Underrated outcome of @AnthropicAI Claude having up to 1m token context length with ~perfect recall: Claude can now confidently handle >250 tools!! When OpenAI fn calling first came out I think it was @simonw who called out that we needed "deep" functions bc of the context
-
Claude Tool Use Now Available Beta Anthropic Messages API
By
–
Tool use is now available in beta to all customers in the Anthropic Messages API, enabling Claude to interact with external tools using structured outputs. pic.twitter.com/kJHhSvyLdK
— Anthropic (@AnthropicAI) 4 avril 2024Tool use is now available in beta to all customers in the Anthropic Messages API, enabling Claude to interact with external tools using structured outputs.