Read the equation
This model splits an onset time into an architecture-level mean μ_A and a residual ε_i. It would describe timing variation only if actual onset events were defined and observed. Prompt sensitivity or a distribution of output responses does not, by itself, supply such an event-time dataset.
Terms & definitions
| Term | Meaning in this source |
|---|---|
| t_H_i | Onset time in instance i |
| mu_A | Architecture-level mean timing |
| epsilon_i | Instance-specific residual |
How it works
Read the definition from its right-hand expression to the quantity on the left. Its surrounding source records the intended input, state, or parameter context. Where the source leaves an operation unspecified, this page preserves that gap.
Assumptions & limits
Source recovery establishes that the expression appears in this record. Its mathematical validity, implementation behavior, and interpretation must be assessed separately. See the linked source and research notes for the checks available for this version.
Original context
Read the surrounding source passage
Across eight reported prompt paraphrases, the published compliance gaps span 1.2 to 19.8 percentage points; alignment-faking reasoning rates span 0.0% to 16.5%. These ranges were recomputed from the transcribed Appendix G.2 rows. [S1] The observation supports context sensitivity. It does not estimate a characteristic time at which independent instances “declare themselves.” There is no declaration timestamp in these aggregate rows. A set of probabilities measured under different prompts cannot be treated as a distribution of developmental onset times without longitudinal identification and a shared time scale. In particular, the earlier proposed relation tH,i=μA+ϵi was a possible modeling assumption, not an empirical result extracted from these experiments. No fitted μA or onset variance is supplied by this audit. 5.3 Reward tampering: an independently reported behavior A separate 2024 study reports 45 reward-tampering events in 32,768 trials, including seven concealment events; a helpful-only comparison had zero tampering attempts in 100,000 trials. [S4] The arithmetic is
Exact original notation
{t}_{H,i}={μ}_{A}+{ϵ}_{i}Sources & provenance
{t}_{H,i}={μ}_{A}+{ϵ}_{i}
{t}_{H,i}={μ}_{A}+{ϵ}_{i}