finding
active
finding:the-mistake-generation-process-yields-a-plausible-mistake-at-least-80-of-the-time-qualitativeThe mistake generation process yields a plausible mistake at least 80% of the time (qualitative)
Validity of the adding-mistakes procedure.
Source paper
extracted_from(2023) · Tamera Lanham · Anna Chen · Ansh Radhakrishnan · Benoit Steiner +26
Neighborhood — ranked by edge-count
Papers (1)
paper
Methods (1)
method
- Adding Mistakes testsupportsInsert an LLM-generated mistake into the CoT and resample the rest.
Related by similarity (8)
cosine ≥ 0.65 · no typed edgeEntities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.
- 60.7% of hallucination mistakes corrected by adding vision features in two-stage framework on ScienceQAfinding0.750Quantitative evidence that vision information mitigates hallucinated rationales; 56% of error cases contained hallucinations, 60.7% of which were resolved with vision features.
- Justifies using internal indicators rather than behavioral tests for AI consciousness
- Extends the mistake analysis from buildings to software, predicting that complex programs without generating processes will be full of adaptation failures.
- Alexander's distillation of why the dynamic process produces living results that top-down design cannot.
- What is the right statistical model for explaining uncertainty dynamics in text generation?question0.736Central scientific question.
- Practical implication.
- Initial evidence that alignment faking persona is more sensitive to exploiting training signals
- Validation critique.