REVIEW 4 cited by
The Curious Case of Hallucinations in Neural Machine Translation
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
In this work, we study hallucinations in Neural Machine Translation (NMT), which lie at an extreme end on the spectrum of NMT pathologies. Firstly, we connect the phenomenon of hallucinations under source perturbation to the Long-Tail theory of Feldman (2020), and present an empirically validated hypothesis that explains hallucinations under source perturbation. Secondly, we consider hallucinations under corpus-level noise (without any source perturbation) and demonstrate that two prominent types of natural hallucinations (detached and oscillatory outputs) could be generated and explained through specific corpus-level noise patterns. Finally, we elucidate the phenomenon of hallucination amplification in popular data-generation processes such as Backtranslation and sequence-level Knowledge Distillation.
Forward citations
Cited by 4 Pith papers
-
Calm-Whisper: Reduce Whisper Hallucination On Non-Speech By Calming Crazy Heads Down
Fine-tuning only three decoder attention heads in Whisper-large-v3 on non-speech audio with blank labels cuts non-speech hallucination on UrbanSound8K by 84.5% with less than 0.1% WER loss on LibriSpeech.
-
Investigation of Whisper ASR Hallucinations Induced by Non-Speech Audio
Recurring Whisper hallucinations on non-speech audio can be catalogued and removed via post-processing, cutting word error rate in augmented-speech tests.
-
GOLFer: Smaller LM-Generated Documents Hallucination Filter & Combiner for Query Expansion in Information Retrieval
GOLFer filters hallucinated sentences from small-LM-generated hypothetical documents and reweights the rest into the query, improving retrieval at lower cost than large LLM expansion.
-
Mitigating Knowledge Conflicts in Language Model-Driven Question Answering
On memorized question-answer pairs from KMIR and NQ, bottleneck and prefix adapters trained on entity-swapped contexts let a GPT-2 reader follow the new context most of the time, though no baselines are reported.
Discussion (0). Continue with ORCID to comment.