Hallucinations both in and out of context: An active inference account
Bibliographic record
Abstract
Hallucinations, including auditory verbal hallucinations (AVH), occur in both the healthy population and in psychotic conditions such as schizophrenia (often developing after a prodromal period). In addition, hallucinations can be in-context (they can be consistent with the environment, such as when one hallucinates the end of a sentence that has been repeated many times), or out-of-context (such as the bizarre hallucinations associated with schizophrenia). In previous work, we introduced a model of hallucinations as false (positive) inferences based on a (Markov decision process) formulation of active inference. In this work, we extend this model to include content-to disclose the computational mechanisms behind in- and out-of-context hallucinations. In active inference, sensory information is used to disambiguate alternative hypotheses about the causes of sensations. Sensory information is balanced against prior beliefs, and when this balance is tipped in the favor of prior beliefs, hallucinations can occur. We show that in-context hallucinations arise when (simulated) subjects cannot use sensory information to correct prior beliefs about hearing a voice, but beliefs about content (i.e. the sequential order of a sentence) remain accurate. When hallucinating subjects also have inaccurate beliefs about state transitions, out-of-context hallucinations occur; i.e. their hallucinated speech content is disordered. Note that out-of-context hallucinations in this setting does not refer to inference about context, but rather to false perceptual inference that emerges when the confidence in-or precision of-sensory evidence is reduced. Furthermore, subjects with inaccurate beliefs about state transitions but an intact ability to use sensory information do not hallucinate and are reminiscent of prodromal patients. This work demonstrates the different computational mechanisms that may underlie the spectrum of hallucinatory experience-from the healthy population to psychotic states.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".