chatRater-package

This package provides functions to interact with multiple LLM APIs
(e.g., 'OpenAI', 'DeepSeek', 'Anthropic', 'Cohere', 'Google PaLM', 'Ollama').
Additionally, several functions are provided that encapsulate LLM prompts to obtain
various psycholinguistic metrics:<ul>
<li>Word Frequency: Number of occurrences (often log-transformed) of a word in a corpus.</li>
<li>Lexical Coverage: Proportion of words in the stimulus that are found in a target vocabulary.</li>
<li>Zipf Metric: The slope of the log-log frequency vs. rank distribution, per Zipf's law.</li>
<li>Levenshtein Distance (D): Minimum number of single-character edits required to transform one stimulus into another.</li>
<li>Semantic Transparency: The degree to which the meaning of a compound or phrase is inferable from its parts.</li>
</ul>

Generates ratings and psycholinguistic metrics for textual stimuli using large language models.
It enables users to evaluate idioms and other language materials by combining context, prompts, and stimulus inputs.
It supports multiple LLM APIs (such as 'OpenAI', 'DeepSeek', 'Anthropic', 'Cohere', 'Google PaLM', and 'Ollama')
by allowing users to switch models with a single parameter. In addition to generating numeric ratings,
'chatRater' provides functions for obtaining detailed psycholinguistic metrics including word frequency (with optional corpus input),
lexical coverage (with customizable vocabulary size and test basis), Zipf metric, Levenshtein distance, and semantic transparency.

Shiyang Zheng

chatRater-package: Multi-Model LLM API Wrapper and Cognitive Experiment Utilities

Description

Arguments

Details