Hallucinations arise from biased latent inference paths rather than missing knowledge, demonstrated via a new diagnostic testbed TrapQA that isolates task-retrieval and key-selection biases.
On the Origin of Hallucinations in Conversational Models: Is it the Datasets or the Models?
4 Pith papers cite this work, alongside 109 external citations. Polarity classification is still indexing.
4
Pith papers citing it
109
external citations · OpenAlex
fields
cs.CL 4representative citing papers
Adapts multi-layer token-level Mahalanobis distance with supervised linear regression to yield improved uncertainty scores for LLM truthfulness tasks.
StarCoderBase matches or beats OpenAI's code-cushman-001 on multi-language code benchmarks; the Python-fine-tuned StarCoder reaches 40% pass@1 on HumanEval while retaining other-language performance.
Hybrid entropy-uncertainty-geometric defence improves clean accuracy by up to 43% and adversarial robustness by up to 65% on NLU and security benchmarks.
citing papers explorer
-
Understanding Why Language Models Hallucinate: Testing Reasoning Against Priors
Hallucinations arise from biased latent inference paths rather than missing knowledge, demonstrated via a new diagnostic testbed TrapQA that isolates task-retrieval and key-selection biases.
-
Token-Level Density-Based Uncertainty Quantification Methods for Eliciting Truthfulness of Large Language Models
Adapts multi-layer token-level Mahalanobis distance with supervised linear regression to yield improved uncertainty scores for LLM truthfulness tasks.
-
StarCoder: may the source be with you!
StarCoderBase matches or beats OpenAI's code-cushman-001 on multi-language code benchmarks; the Python-fine-tuned StarCoder reaches 40% pass@1 on HumanEval while retaining other-language performance.
-
Hybrid Adversarial Defence for Natural Language Understanding Tasks
Hybrid entropy-uncertainty-geometric defence improves clean accuracy by up to 43% and adversarial robustness by up to 65% on NLU and security benchmarks.