A collection of LogitsProcessors to customize and enhance LLM behavior for specific tasks.
-
Updated
Jul 8, 2025 - Python
A collection of LogitsProcessors to customize and enhance LLM behavior for specific tasks.
Plots how the logit values that are passed into the softmax function change over time as the model is trained.
Novel Hallucination detection method
Detecting prompt toxicity from a small LLM's internal logits and embeddings, no separate moderation model needed. 99.18% AUPRC with a 64-neuron MLP.
Temporal analysis of LLM safety activation via logit-margin scores.
Threshold determination app hosted with Streamlit to determine optimal classification thresholds for prediction algorithms.
This project bridges the gap between human language and machine-executable code by translating prompts into structured function calls with typed arguments. Built strictly with Python 3.10+ and Pydantic , it demonstrates how constrained decoding can ensure near-perfect reliability and strict JSON compliance, even on small 0.5B parameter models.
Convolutional Networks Training and Logits Saving
Multi strategy logit fusion between Ouro-1.4B (Universal Transformer) and HRM-Text-1B (prefix-LM) for improved text generation
GGUF parity and model-validation tools for comparing logits and hidden states against llama.cpp or other references.
Add a description, image, and links to the logits topic page so that developers can more easily learn about it.
To associate your repository with the logits topic, visit your repo's landing page and select "manage topics."