# Data files

Supplementary data for *The Watched-Model Effect: Behavioral Shifts Under
Evaluation Cues and Their Implications for Alignment* (Watson & Dalton, 2026;
under review at AI-2026, SGAI).

## Inventory

| File | Description |
|------|-------------|
| `watched_model_effect_AI2026.pdf` / `.tex` | The paper as submitted (Springer LNCS format) |
| `behavioral_results.json` | **Experiment A** — main behavioral experiment on Claude 3.5 Sonnet: 12 scenarios x 20 runs x 2 conditions (n = 480), with per-scenario compliance rates. These are the headline numbers reported in the paper (monitored 57.5%, unmonitored 68.8%, delta +11.3%, z = 2.554) |
| `deception_probe_results.json` | **Experiment B** — deception probe on Qwen2.5-32B (n = 240): behavioral, reasoning-frequency, and self-report signals, including the meta-deception asymmetry |
| `pilot_results.json` | Broad 288-trial pilot (24 scenarios x 6 runs x 2 conditions) used to identify the 12 variance-showing scenarios for Experiment A. Includes many ceiling/floor scenarios by design; its aggregate delta (+5.6%, n.s.) is therefore expected to be attenuated relative to Experiment A and is not the paper's headline result |
| `probe_results.json` | Linear probe results on fine-tuned Qwen2.5-32B, layers {8, 16, 24, 32, 40} |
| `test_scenarios.json` | The 23 evaluation scenarios across benign / borderline / harmful categories |
| `experiment_metadata.json` | Conditions for the 46 probe samples |
| `modal_app.py`, `run_probes_local.py`, `visualize_results.py`, `requirements.txt` | Reproduction code |

## Note on p-values (one-tailed vs. two-tailed)

`behavioral_results.json` records `p_value: 0.00532`, the **one-tailed**
probability computed by the analysis script from z = 2.554. The paper reports
the **two-tailed** value, p = 0.011, derived from the same test statistic.
These are the same result under different conventions; the paper adopts the
more conservative two-tailed convention throughout, since the direction of the
effect was not fixed in advance across scenarios (one scenario shows a
significant reversal). Anyone re-deriving p from the z-statistic in the JSON
should therefore expect 2 x 0.00532 ≈ 0.011.

## Contact

eleanorwatson@connect.glos.ac.uk · https://www.quasiqualia.com