Log probabilities (logprobs) provide a window into how confident a language model is about its predictions. In this technical implementation, we demonstrate how to access and interpret logprobs via the OpenAI API, using a series of increasingly difficult multiplication tasks. Our experiment reveals that declining confidence scores can effectively signal
Tool use (also called function calling) is one of the most powerful capabilities available in modern language models. It allows models to extend beyond text generation by interacting with external systems, databases, or custom logic—making AI agents capable of real-world tasks like checking weather, querying APIs, or running calculations.
In our previous experiment, we showed how DeepSeek-V3 could classify argumentative claims in a simple prompt-based setup. In this follow-up, we take the test one step further: running the same PRO/CON stance classification task programmatically via the DeepSeek API in a Google Colab environment. This shift from manual prompting
Can state-of-the-art multimodal models analyse animal behaviour directly from video footage? In this study, we tested Google’s Gemini 2.5 Pro — both in AI Studio and via its API — to assess whether it can produce structured ethological descriptions based purely on short animal-related videos. By applying a consistent prompt
We assessed the ability of Claude 4 Sonnet and GPT-4o to classify brain MRI images as healthy or tumorous for research labelling purposes, using both the chat interface (10 images) and the API (125 images). Claude 4 Sonnet achieved perfect accuracy (10/10) in chat, but its API refused to
Accessing large language models (LLMs) through APIs opens up research and development opportunities that go well beyond the limits of traditional chat interfaces. For researchers API access enables language models to be built directly into bespoke workflows, analytical tools, or automated processes—supporting tasks such as large-scale text analysis, data
Image-based classification is increasingly used across biology, ecology, and agriculture—from identifying animal species to detecting plant diseases. One common use case is the analysis of leaf images to distinguish between healthy and diseased plants. In this post, we compare two approaches to classifying strawberry leaves as either fresh (healthy)
This post presents a practical solution for integrating OpenAI’s GPT API into RStudio using a custom Shiny interface. The tool enables real-time code generation from natural language instructions, allowing users to interact with GPT-4 directly within their R workflow—without leaving the environment or blocking the console. We demonstrate
Large language models (LLMs) rely on a set of parameters that directly influence how text is generated — affecting randomness, repetition, length, and coherence. Understanding of these parameters is essential when working with LLMs in research, application development, or evaluation settings. While chat-based interfaces such as ChatGPT, Copilot, or Gemini typically
When working with scanned documents, having direct access to a reliable OCR API can make text extraction faster, more consistent, and easier to integrate into existing workflows. In a previous post, we explored a prompt-based solution using Mistral’s Le Chat interface—which produced clean results but missed certain elements
