6 papers
There Is No Neutral Harness: Modern LLM Leaderboards Are Manufactured by Config-Fragile Items
V. S. Raghu Parupudi
Multiple-choice benchmarks fix the questions and the correct answers, but not the harness: the order of the options, the wording of the prompt, and whether a language model's answe…
Anisotropy Decides Cosine vs. Rank Metrics for Text Embeddings
V. S. Raghu Parupudi
The standard way to compare two text embeddings is cosine similarity. Scattered studies report that a different metric does better, but never pin down the geometric condition that…
Before and After Temperature: A Distributional View of Creative LLM Generation
V. S. Raghu Parupudi, Harsha Ponnada, Aditi Kaushal +3
Reference-free evaluation of large language model (LLM) creativity relies on perplexity, entropy, and top-1 margin. We show that a much stronger signal lives one step earlier in th…
Confidence, Not Perplexity: A Better Metric for the Creative Era of LLMs
V. S. Raghu Parupudi
Reference-free metrics like self-perplexity are strongly biased against creative text generation. We propose the Confidence Score (CS), derived from a model's output probability di…
Systematic Diagnosis of Brittle Reasoning in Large Language Models
V. S. Raghu Parupudi
A central question in artificial intelligence is the extent to which machine learning models comprehend mathematics. To address this, we propose a novel framework for measuring mat…
Magnitude Matters: a Superior Class of Similarity Metrics for Holistic Semantic Understanding
V. S. Raghu Parupudi
Vector comparison in high dimensions is a fundamental task in NLP, yet it is dominated by two baselines: the raw dot product, which is unbounded and sensitive to vector norms, and…