Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2509.11208 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866910028469895168 |
|---|---|
| author | Chlon, Leon Karim, Ahmed Chlon, Maggie Awada, MarcAntonio |
| author_facet | Chlon, Leon Karim, Ahmed Chlon, Maggie Awada, MarcAntonio |
| contents | Transformers used for evidence-grounded question answering with binary adjudication (e.g., support/refute or yes/no) can be highly sensitive to the order in which exchangeable evidence is presented, producing dispersion across permutations and unreliable attempted answers (``hallucinations'' under a Bernoulli predicate).
We treat evidence order as a nuisance variable and show that next-token training minimizes expected conditional description length over orderings. This objective can be close to Bayes-optimal in expectation while deviating under any fixed ordering. We quantify this expectation--realization gap via a Quantified Martingale Violation (QMV) bound that predicts $\mathcal{O}(\log n)$ growth in permutation dispersion under harmonic positional sensitivity.
We then derive the Expectation-level Decompression Law (EDFL), relating expected information budget to achievable reliability for Bernoulli predicates, and use it to define \emph{Bits-to-Trust} (B2T), \emph{Risk-of-Hallucination} (RoH), and the \emph{Information Sufficiency Ratio} (ISR), together with a fixed ISR-gating rule for answer/abstain decisions under permutation mixtures.
On 3,059 grounded items from a five-benchmark evidence-grounded QA suite (FEVER, HotpotQA, NQ-Open, PopQA, and Controls), we observe logarithmic dispersion and Jensen gains from uniform permutation mixtures. In a pre-specified held-out audit (528 items), an ISR $= 1$ gate attains 0.0--0.7\% hallucination with 20.6--27.9\% abstention (95\% confidence intervals). |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2509_11208 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Predictable Compression Failures: Order Sensitivity and Information Budgeting for Evidence-Grounded Binary Adjudication Chlon, Leon Karim, Ahmed Chlon, Maggie Awada, MarcAntonio Machine Learning Transformers used for evidence-grounded question answering with binary adjudication (e.g., support/refute or yes/no) can be highly sensitive to the order in which exchangeable evidence is presented, producing dispersion across permutations and unreliable attempted answers (``hallucinations'' under a Bernoulli predicate). We treat evidence order as a nuisance variable and show that next-token training minimizes expected conditional description length over orderings. This objective can be close to Bayes-optimal in expectation while deviating under any fixed ordering. We quantify this expectation--realization gap via a Quantified Martingale Violation (QMV) bound that predicts $\mathcal{O}(\log n)$ growth in permutation dispersion under harmonic positional sensitivity. We then derive the Expectation-level Decompression Law (EDFL), relating expected information budget to achievable reliability for Bernoulli predicates, and use it to define \emph{Bits-to-Trust} (B2T), \emph{Risk-of-Hallucination} (RoH), and the \emph{Information Sufficiency Ratio} (ISR), together with a fixed ISR-gating rule for answer/abstain decisions under permutation mixtures. On 3,059 grounded items from a five-benchmark evidence-grounded QA suite (FEVER, HotpotQA, NQ-Open, PopQA, and Controls), we observe logarithmic dispersion and Jensen gains from uniform permutation mixtures. In a pre-specified held-out audit (528 items), an ISR $= 1$ gate attains 0.0--0.7\% hallucination with 20.6--27.9\% abstention (95\% confidence intervals). |
| title | Predictable Compression Failures: Order Sensitivity and Information Budgeting for Evidence-Grounded Binary Adjudication |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2509.11208 |