synthetic-hallucination-benchmarks-where-faithful-items-human-written-hallucinated-itemsIn synthetic hallucination benchmarks where faithful items are human-written and hallucinated items are LLM-rewrites, a detector's binary score partly measures human-vs-machine style rather than factuality: feeding faithful LLM rewrites of the same truthful articles makes some fine-tuned encoders flag nearly all of them as hallucinated.
Ingested from a paper but not yet reviewed by a human. It is deliberately inert: it does not move any technique’s standing, does not count toward the backtest, and is excluded anywhere a claim would carry weight. Read the source before relying on it.
Capability: Whether the measurement made the finding
Observed on
Fine-tuned multilingual encoders (ReMBERT, mDeBERTa) trained on an LLM-generated news hallucination corpus; provenance analysis run on Russian original text only..
Sources
- Measured on 290 truthful human articles, 290 factcheck.kz human-written fakes, 290 faithful Gemini rewrites, and synthetic hallucinations. ReMBERT flagged 0.352 of human truthful vs 1.000 of faithful LLM rewrites; mDeBERTa 0.221 vs 0.586, so the effect is real but model-dependent. Single language, single domain, two encoders; register differences between real fakes and newswire are acknowledged but not fully controlled.
Disagreeing is the most useful thing you can do here. Both sides of every contested claim in this catalog were assembled by the same person, which is its weakest point.
Related claims
- For training-free inference-time hallucination mitigation in 7B vision- language models, reported gains on object-hallucination benchmarks are largely inseparable from reduced informativeness: hallucination rate and object recall/coverage move together, so a lower score can mean the model mentioned fewer visual entities rather than grounded better.Whether the measurement made the finding · unreviewed
- General-purpose factuality verifiers (entailment-, consistency-, and RAG-trained) and zero-shot frontier LLM judges perform near chance at spotting unsupported claims in scientific peer reviews that must be checked against the full submitted paper, because they cannot separate paper-unsupported assertions from legitimate evaluative critique; small models fine-tuned on in-domain examples beat them by a wide margin.Stating false facts confidently · unreviewed
- For detecting hallucinated text spans in vision-language model outputs, span localization and confidence calibration come from different architectures: a fine-tuned generative VLM proposes better span boundaries but scores them poorly, while a discriminative token tagger trained with a direct regression loss on human-consensus confidence gives better-correlated probabilities but conservative spans, so using the generative spans as proposals and the tagger's probabilities as scores beats either alone.Stating false facts confidently · unreviewed
- For black-box hallucination detection using only sampled responses and token log-probabilities, semantic-entropy signals go to zero whenever the samples collapse into one semantic cluster — which happens for a large share of hallucinated queries, up to nearly all of them on tasks with confusable entities and figures — while token-level uncertainty still separates many of those cases, so the two signals fail on different queries and combining them detects more than either alone.Stating false facts confidently · unreviewed
- In LLM-based ASR, steering decoding toward acoustic evidence only helps if candidates are first restricted to tokens within a small likelihood margin of the greedy token: unconstrained selection by an acoustic compatibility score removes hallucinated outputs but destroys transcription accuracy through unbounded insertions, while the constrained version removes a large share of hallucinations at near-zero WER/CER cost.Stating false facts confidently · unreviewed
Notes
Drafted from the paper by a model and filed unreviewed. Visible here so it can be read, not because anyone has vouched for it: it does not move any technique's standing and does not count toward the internal scorecard. Drafted confidence: medium. Falsifier as drafted: Detectors trained on such synthetic splits show near-baseline flag rates on faithful LLM rewrites of truthful articles, i.e. their scores track veracity independently of text provenance. Proposed technique, not catalogued: Provenance-controlled evaluation split.