Langkit

🔍 LangKit: An open-source toolkit for monitoring Language Learning Models (LLMs). 📚 Extracts signals from prompts & responses, ensuring safety & security. 🛡️ Features include text quality, relevance metrics, & sentiment analysis. 📊 A comprehensive tool for LLM observability. 👀
This version includes:

- Add vader_sentiment support 218 [jamie256]
- deprecate OpenAIDavinci and add OpenAILegacy 216 [FelipeAdachi]
- Add PII module 217 [FelipeAdachi]
- Injections example - replacing deprecated openai model 214 [FelipeAdachi]
- Adds example segmenting on a validation column produced post-UDF 153 [richard-rogers]
- Jailbreak and Prompt Injection Example Notebook 211 [FelipeAdachi]
- Bump version to 0.0.28 210 [w0-automator]

🔍 LangKit: An open-source toolkit for monitoring Language Learning Models (LLMs). 📚 Extracts signals from prompts & responses, ensuring safety & security. 🛡️ Features include text quality, relevance metrics, & sentiment analysis. 📊 A comprehensive tool for LLM observability. 👀
This version includes:

- Update description, minor cleanup of README.md and tests 209 [jamie256]
- Openai client compatibility with both v0 and v1 208 [FelipeAdachi]
- Proactive Injection Detection 201 [FelipeAdachi]
- better prompt for llm consistency check 206 [FelipeAdachi]
- add extract scenario to Logging_Text.ipynb 203 [FelipeAdachi]
- Bump version to 0.0.27 202 [w0-automator]

🔍 LangKit: An open-source toolkit for monitoring Language Learning Models (LLMs). 📚 Extracts signals from prompts & responses, ensuring safety & security. 🛡️ Features include text quality, relevance metrics, & sentiment analysis. 📊 A comprehensive tool for LLM observability. 👀
This version includes:

- Feature extraction - wrapper around schema.apply_udf 198 [FelipeAdachi]
- CUDA benchmarks 194 [richard-rogers]
- Bump version to 0.0.26 200 [w0-automator]

🔍 LangKit: An open-source toolkit for monitoring Language Learning Models (LLMs). 📚 Extracts signals from prompts & responses, ensuring safety & security. 🛡️ Features include text quality, relevance metrics, & sentiment analysis. 📊 A comprehensive tool for LLM observability. 👀
This version includes:

- Add support for Azure OpenAI hosted models 199 [jamie256]
- Control GPU usage in injections module 193 [richard-rogers]
- Bump version to 0.0.25 192 [w0-automator]

🔍 LangKit: An open-source toolkit for monitoring Language Learning Models (LLMs). 📚 Extracts signals from prompts & responses, ensuring safety & security. 🛡️ Features include text quality, relevance metrics, & sentiment analysis. 📊 A comprehensive tool for LLM observability. 👀
This version includes:

- Optionally disable CUDA for Encoder metric modules 191 [richard-rogers]
- Update release workflow to use whylabs automator 190 [jamie256]
- Add conditional CUDA support with env variable 189 [richard-rogers]
- Preliminary benchmarking 185 [richard-rogers]
- Bump version: 0.0.23 → 0.0.24 187 [jamie256]

🔍 LangKit: An open-source toolkit for monitoring Language Learning Models (LLMs). 📚 Extracts signals from prompts & responses, ensuring safety & security. 🛡️ Features include text quality, relevance metrics, & sentiment analysis. 📊 A comprehensive tool for LLM observability. 👀
This version includes:

- Add langkit metadata to schema for inclusion in whylogs profiles 186 [jamie256]
- security features doc update to include hallucinations 182 [FelipeAdachi]
- Link reference paper for Hallucinations metric module doc. 181 [jamie256]
- Add hallucinations to README and update modules list to fix formatting 180 [jamie256]
- add hallucination docs 179 [FelipeAdachi]
- Bump version: 0.0.22 → 0.0.23 183 [jamie256]