How accurate and predictive are judgments of solvability? Explorations in a two-phase anagram solving paradigm
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
Abstract Meta-reasoning requires monitoring and controlling one’s reasoning processes, and it often begins with an assessment of problem solvability. We explored whether Judgments of Solvability (JOS) for solvable and unsolvable anagrams discriminate and predict later problem-solving outcomes once anagrams solved during the JOS task are excluded. We also examined whether providing training via longer-duration anagrams improves JOS discrimination and predictiveness. In a two-phase paradigm, participants judged each anagram as solvable , not solvable , or already solved ( S , NS , AS ; JOS phase ) then later attempted to solve the anagrams within 45 s ( solving phase ). Anagrams were presented in 4 blocks. In the training groups , anagram duration started at 16 s and halved across blocks, whereas in the no-training groups anagram duration was always 2 s. Participants’ S JOSs typically were discriminating after excluding anagrams that received AS JOSs, but training did not lead to better discrimination in the final block. Training improved AS JOS predictiveness, but not S JOS predictiveness. Thus, training increased solving during the JOS task rather than increasing JOS predictiveness. In Experiment 3 these findings replicated when both solvable and unsolvable anagrams were presented in the solving phase and no response deadline was set. Here, problem-solving outcomes and effort regulation (i.e., response times) were predicted by AS and NS JOSs, but not by S JOSs. Overall, although S JOSs were discriminating, they were not predictive of later problem solving or effort regulation—and this was true even after training with longer-duration anagrams.
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.
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 | 0.003 | 0.002 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.001 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 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 it