REVIEW 9 cited by
The Troubling Emergence of Hallucination in Large Language Models -- An Extensive Definition, Quantification, and Prescriptive Remediations
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
Signed reviews
read the original abstract
The recent advancements in Large Language Models (LLMs) have garnered widespread acclaim for their remarkable emerging capabilities. However, the issue of hallucination has parallelly emerged as a by-product, posing significant concerns. While some recent endeavors have been made to identify and mitigate different types of hallucination, there has been a limited emphasis on the nuanced categorization of hallucination and associated mitigation methods. To address this gap, we offer a fine-grained discourse on profiling hallucination based on its degree, orientation, and category, along with offering strategies for alleviation. As such, we define two overarching orientations of hallucination: (i) factual mirage (FM) and (ii) silver lining (SL). To provide a more comprehensive understanding, both orientations are further sub-categorized into intrinsic and extrinsic, with three degrees of severity - (i) mild, (ii) moderate, and (iii) alarming. We also meticulously categorize hallucination into six types: (i) acronym ambiguity, (ii) numeric nuisance, (iii) generated golem, (iv) virtual voice, (v) geographic erratum, and (vi) time wrap. Furthermore, we curate HallucInation eLiciTation (HILT), a publicly available dataset comprising of 75,000 samples generated using 15 contemporary LLMs along with human annotations for the aforementioned categories. Finally, to establish a method for quantifying and to offer a comparative spectrum that allows us to evaluate and rank LLMs based on their vulnerability to producing hallucinations, we propose Hallucination Vulnerability Index (HVI). We firmly believe that HVI holds significant value as a tool for the wider NLP community, with the potential to serve as a rubric in AI-related policy-making. In conclusion, we propose two solution strategies for mitigating hallucinations.
Forward citations
Cited by 9 Pith papers
-
Neural Message-Passing on Attention Graphs for Hallucination Detection
CHARM trains graph neural networks on token-attention graphs built from LLM computational traces and outperforms prior hallucination detectors on five benchmarks at token and response level.
-
LLM Embedding-based Attribution (LEA): Quantifying Source Contributions to Generative Model's Response for Vulnerability Analysis
LEA uses rank-based linear dependence of layer-0 hidden states to attribute each response token to query, retrieved context, or internal knowledge, and distinguishes valid from generic retrieval with over 95% accuracy.
-
Probe-Free Low-Rank Activation Intervention
FLORAIN is a probe-free, single-layer activation intervention that improves LLM truthfulness by projecting hidden states toward an ellipsoidal region of desirable answers.
-
Position: Stop Acting Like Language Model Agents Are Normal Agents
LMAs should be treated as systems whose agency is unstable and must be measured, not as normal agents with persistent identity.
-
Loki's Dance of Illusions: A Comprehensive Survey of Hallucination in Large Language Models
A survey of LLM hallucination research that formalizes hallucination types and argues, via incompleteness and undecidability arguments, that hallucinations cannot be fully eliminated.
-
ChallengeMe: An Adversarial Learning-enabled Text Summarization Framework
ChallengeMe applies an adversarial generation-evaluation-feedback prompt loop to LLM summarization, but its claimed superiority over GPT-4o and other baselines is undermined by test-set threshold tuning and internal c...
-
Self-Disclosure to AI: The Paradox of Trust and Vulnerability in Human-Machine Interactions
People disclose personal information to AI because it seems non-judgmental, and this same perception creates privacy and emotional risks that current ethics frameworks do not address.
-
Future Research Avenues for Artificial Intelligence in Digital Gaming: An Exploratory Report
An exploratory report curating five promising AI-for-gaming research avenues, with no original findings.
-
Emerging Security Challenges of Large Language Models
A workshop report summarizes adversarial attack surfaces of LLMs, from training data poisoning to prompt injection, and calls for more research on defenses.
Discussion (0). Continue with ORCID to comment.