What question did this study set out to answer?

To enhance the Livnium NLI system by integrating multi-scale attractor dynamics for better inference accuracy.

April 8, 20260 citationsOpen Access

Livnium v3: Multi-Scale Attractor Dynamics, Token-Level Alignment, and Divergence-Based Reliability for Natural Language Inference

Key Points

To enhance the Livnium NLI system by integrating multi-scale attractor dynamics for better inference accuracy.
Introduced a cross-encoder framework to improve accuracy.
Extracted token-level alignment using BERT's cross-attention layer.
Developed a divergence metric to assess prediction reliability.
Achieved an increase in SNLI development accuracy from 82.2% to 84.5%.
Provided a force map visualization for structural comparisons between tokens.
Established thresholds for differentiating reliable and unreliable predictions.

Abstract

We extend the Livnium attractor-based NLI system with three contributions: (1) a cross-encoder upgrade improving SNLI dev accuracy from 82.2% to 84.5% via joint CLS premise SEP hypothesis SEP encoding; (2) token-level alignment extraction from the last-layer BERT cross-attention block, rendering the model's internal structural comparison visible as a force map between premise and hypothesis tokens; and (3) an alignment divergence metric that serves as a zero-cost intrinsic reliability signal, with empirically validated thresholds distinguishing stable from unreliable predictions. We further demonstrate that the same constraint-injection mechanism reproduces the Bayesian belief update in the Monty Hall problem, connecting NLI inference to classical decision theory through a unified energy-reshaping framework.

Ask AI

Helpful

Bookmark

View Full Paper