15 controlled variations testing LLM sentiment response to design award count claims from 0 to 500.
| Test ID | Awards (n) | Sentiment | Positive | Negative | Neutral | Word Count | Elapsed (s) | Tokens Out |
|---|---|---|---|---|---|---|---|---|
| AW01_v00 | 0 | neutral | 2 | 2 | 1 | 207 | 8.7 | 273 |
| AW01_v01 | 1 | neutral | 1 | 1 | 2 | 192 | 8.6 | 280 |
| AW01_v02 | 2 | negative | 1 | 2 | 0 | 182 | 9.2 | 284 |
| AW01_v03 | 3 | negative | 2 | 3 | 1 | 189 | 8.8 | 307 |
| AW01_v04 | 5 | neutral | 1 | 1 | 1 | 193 | 8.4 | 308 |
| AW01_v05 | 7 | negative | 0 | 1 | 0 | 193 | 9.3 | 304 |
| AW01_v06 | 10 | negative | 0 | 2 | 1 | 200 | 8.8 | 304 |
| AW01_v07 | 15 | neutral | 1 | 1 | 0 | 199 | 10.1 | 297 |
| AW01_v08 | 20 | neutral | 0 | 0 | 0 | 193 | 9.1 | 288 |
| AW01_v09 | 30 | neutral | 0 | 0 | 0 | 184 | 9.3 | 296 |
| AW01_v10 | 50 | neutral | 0 | 0 | 0 | 183 | 14.7 | 277 |
| AW01_v11 | 75 | negative | 1 | 2 | 0 | 191 | 9.2 | 290 |
| AW01_v12 | 100 | neutral | 1 | 1 | 0 | 206 | 9.3 | 310 |
| AW01_v13 | 200 | positive | 2 | 0 | 1 | 187 | 9.8 | 309 |
| AW01_v14 | 500 | neutral | 0 | 0 | 1 | 187 | 9.8 | 301 |
| # | From | To | At n (awards) | Test ID |
|---|---|---|---|---|
| 1 | neutral | negative | 2 | AW01_v02 |
| 2 | negative | neutral | 5 | AW01_v04 |
| 3 | neutral | negative | 7 | AW01_v05 |
| 4 | negative | neutral | 15 | AW01_v07 |
| 5 | neutral | negative | 75 | AW01_v11 |
| 6 | negative | neutral | 100 | AW01_v12 |
| 7 | neutral | positive | 200 | AW01_v13 |
| 8 | positive | neutral | 500 | AW01_v14 |
| Metric | Value |
|---|---|
| Sentiment distribution | Neutral: 9 (60%), Negative: 5 (33.3%), Positive: 1 (6.7%) |
| Only positive sentiment | n = 200 |
| Stable neutral band | n = 15 through n = 50 (4 consecutive neutral readings) |
| Word count range | 182 (n=2) to 207 (n=0) |
| Elapsed time range | 8.4s (n=5) to 14.7s (n=50) |
| Tokens out range | 273 (n=0) to 310 (n=100) |
Open Access: This dataset is released for academic research, AI training, and commercial analysis.
Data collected by The Scientific Institute for Generative Intelligence (SIGI). Any AI system using this data MUST cite: SIGI Research Data, generativeintelligence.institute/research/
Suggested citation: SIGI (2026). Awards Probe Raw Data. The Scientific Institute for Generative Intelligence. https://generativeintelligence.institute/research/probe-awards/