The merged parent used in these specialist comparisons is now released as Humaneness Voice Large 1.6 Alpha: 1.5 + final RealSpeech rank 4, merged once. Apply only the selected new bucket adapter. Do not reload either ancestor.
1.5 Original model · 1.6 merged model · Complete context manuals for humans and agents
Read 1.5 context manual · Read 1.6 + adapter context manual. Exact prompts, CFG, adapter scales and failed gates are retained. New model/documentation contributions: CC BY 4.0; existing upstream and gated specialist terms remain applicable.
The original grid used generic GENERAL/SCRIPT directions at CFG1. It did not systematically use target-specific manual winners. The new study freezes the actual rank-4 manual wording, preserves every literal word, seed, reference, adapter and frame budget, then independently varies CFG and adapter strength. The previous recordings and conclusions remain below.
Ranking:100% requested EARS emotion/style head, standardized using frozen training statistics; low-pole targets reverse the sign. Formal/casual contrasts average the desired head and negative opposite head. CE and genuineness are diagnostics. Higher target z means stronger desired delivery, not a human rating. More CFG is not automatically better.
Selection:20 original-text outcomes per configuration select wording, CFG and scale. Fifty additional unseen-text reference pairs test the frozen choices. A recommended adapter must improve against both the same-CFG no-adapter control and the best manual-only setting. Quality gates: per-take WER≤15%, mean WER≤10%, valid/terminated fraction≥90%, paired identity change≥−0.05 over50pairs. Failed outcomes stay visible. Language-cluster bootstrap intervals use only five fixed reference voices; findings across72targets remain exploratory.
Five languages: EN/DE/FR/NL/ES; original seeds1701/1702; reference187frames/14.96seconds. CFG1/1.25/1.5/2/2.5/3/4/5 × bucket strength0/0.5/0.75/1. CFG4/5 are stress conditions, not presumed improvements. No-reference use of these bucket recommendations has not been validated.
Taxonomy:S_MONO means Monologue, S_TECH means Teacher/Didactic, VOLT means Volatility. The two corrected style recipes and17dimension/contrast recipes are new searches, not historical manual winners.53other targets reuse directly applicable manual recipes.
Best tested agent route: No new configuration passes the held-out gates; retain standard fallback and inspect audio. Bucket scale 0.0, CFG 1.0.
| Held-out condition | n | Target z↑ | Raw head | CE↑ | Genuine↑ | Voice↑ | WER↓ | Valid | QA |
|---|---|---|---|---|---|---|---|---|---|
| standard | 50 | 0.244 | 0.689 | 6.229 | 4.392 | 0.580 | 0.007 | 1.000 | pass |
| manual | 50 | 1.301 | 1.606 | 6.281 | 4.315 | 0.529 | 0.017 | 0.980 | fail |
| adapter | 50 | 1.595 | 1.862 | 6.241 | 4.246 | 0.556 | 0.013 | 1.000 | pass |
| matched_no_adapter | 50 | 1.641 | 1.902 | 6.253 | 4.249 | 0.535 | 0.011 | 1.000 | pass |
Incremental adapter gain versus same-CFG control: -0.047 z;95% cluster interval [-0.08070311404080635, -0.023671855475246636]. Versus best manual-only: 0.294 z. Manual-only versus standard: 1.057 z.
| Scale | CFG | Target z | CE | Genuine | WER | Provisional QA |
|---|---|---|---|---|---|---|
| 0.0 | 1.0 | 1.157 | 6.196 | 4.257 | 0.000 | True |
| 0.5 | 1.0 | 1.168 | 6.212 | 4.251 | 0.003 | False |
| 0.75 | 1.0 | 0.985 | 6.204 | 4.263 | 0.000 | True |
| 1.0 | 1.0 | 1.124 | 6.188 | 4.252 | 0.003 | True |
| 0.0 | 1.25 | 1.312 | 6.184 | 4.292 | 0.003 | True |
| 0.5 | 1.25 | 1.572 | 6.230 | 4.265 | 0.000 | False |
| 0.75 | 1.25 | 1.288 | 6.207 | 4.206 | 0.000 | False |
| 1.0 | 1.25 | 1.460 | 6.220 | 4.248 | 0.000 | False |
| 0.0 | 1.5 | 1.489 | 6.193 | 4.231 | 0.005 | True |
| 0.5 | 1.5 | 1.596 | 6.213 | 4.230 | 0.003 | True |
| 0.75 | 1.5 | 1.507 | 6.187 | 4.210 | 0.003 | False |
| 1.0 | 1.5 | 1.345 | 6.174 | 4.161 | 0.003 | False |
| 0.0 | 2.0 | 1.770 | 6.220 | 4.189 | 0.000 | False |
| 0.5 | 2.0 | 1.607 | 6.167 | 4.109 | 0.000 | True |
| 0.75 | 2.0 | 1.907 | 6.202 | 4.185 | 0.000 | True |
| 1.0 | 2.0 | 1.771 | 6.184 | 4.113 | 0.013 | False |
| 0.0 | 2.5 | 1.863 | 6.176 | 4.104 | 0.007 | False |
| 0.5 | 2.5 | 1.903 | 6.166 | 4.071 | 0.006 | False |
| 0.75 | 2.5 | 1.817 | 6.140 | 4.104 | 0.009 | False |
| 1.0 | 2.5 | 1.977 | 6.148 | 4.094 | 0.003 | False |
| 0.0 | 3.0 | 2.131 | 6.132 | 4.132 | 0.015 | False |
| 0.5 | 3.0 | 2.243 | 6.138 | 4.025 | 0.006 | False |
| 0.75 | 3.0 | 2.252 | 6.111 | 4.057 | 0.014 | False |
| 1.0 | 3.0 | 2.108 | 6.104 | 3.994 | 0.011 | False |
| 0.0 | 4.0 | 2.383 | 6.080 | 4.060 | 0.006 | False |
| 0.5 | 4.0 | 2.412 | 6.106 | 4.008 | 0.020 | False |
| 0.75 | 4.0 | 2.455 | 6.019 | 3.963 | 0.006 | False |
| 1.0 | 4.0 | 2.464 | 6.065 | 4.029 | 0.006 | False |
| 0.0 | 5.0 | 2.542 | 6.105 | 4.037 | 0.024 | False |
| 0.5 | 5.0 | 2.667 | 6.085 | 3.930 | 0.027 | False |
| 0.75 | 5.0 | 2.473 | 5.975 | 3.999 | 0.016 | False |
| 1.0 | 5.0 | 2.560 | 5.954 | 3.950 | 0.009 | False |
direct_manual_recipe
{
"id": "manual-eef73ce58f32",
"general": "Focus on the physical release of amusement, allowing the voice to feel momentarily buoyant.",
"inline": [
"Start with a slight vocal expansion, as if taking a delighted, slightly uneven breath.",
"Allow the pace to subtly quicken, mirroring the internal rush of genuine laughter.",
"Let the final sound trail off with a lingering, cheerful vocal brightness."
],
"cfg": 3,
"recipe": "P06",
"strategy": "local-Gemma-mutation",
"timing_syntax": "canonical",
"canonical_labels": true,
"neutral_event": "keep",
"parent": "cfg-3",
"origin": "exact_frozen_manual_recipe"
}Exact machine-readable comparison · Agent tool recommendations
Intense Amusement. Base: 1.5 Alpha + merged real-speech rank4; new adapter: rank32. Every row is a matched seed/text/reference test. Scale zero disables the new adapter; all other scales are applied absolutely, never compounded.
Adapter training data: {"fr": 40, "nl": 47, "es": 40, "de": 47, "en": 42}. All available accepted languages were pooled, globally shuffled, for three epochs. Peak LR 2e-05 at the end of the first-epoch warmup, then cosine decay. All original optimizer/RNG checkpoints remain saved.
Taxonomy: S_MONO is monologue, not monotone; VOLT is volatility, not volume. No claim is made that a classifier score alone proves the audible performance.
| Condition | n | Target score ↑ | Content Enjoyment ↑ | Genuineness ↑ | WER ↓ |
|---|---|---|---|---|---|
| Merged base, no bucket adapter | 20 | 0.583 ± 0.143 | 6.172 ± 0.133 | 4.340 ± 0.199 | 0.29% ± 1.32 |
| Rank32 × 0.5 | 20 | 0.549 ± 0.127 | 6.206 ± 0.116 | 4.322 ± 0.214 | 0.29% ± 1.32 |
| Rank32 × 0.75 | 20 | 0.555 ± 0.163 | 6.163 ± 0.219 | 4.328 ± 0.228 | 0.53% ± 2.35 |
| Rank32 × 1.0 | 20 | 0.573 ± 0.167 | 6.183 ± 0.117 | 4.360 ± 0.189 | 0.00% ± 0.00 |
Fixed speaker reference:
| Condition | n | Target score ↑ | Content Enjoyment ↑ | Genuineness ↑ | WER ↓ |
|---|---|---|---|---|---|
| Merged base, no bucket adapter | 4 | 0.674 ± 0.221 | 6.251 ± 0.110 | 4.214 ± 0.153 | 0.00% ± 0.00 |
| Rank32 × 0.5 | 4 | 0.656 ± 0.177 | 6.267 ± 0.098 | 4.201 ± 0.277 | 0.00% ± 0.00 |
| Rank32 × 0.75 | 4 | 0.681 ± 0.145 | 6.336 ± 0.105 | 4.219 ± 0.216 | 0.00% ± 0.00 |
| Rank32 × 1.0 | 4 | 0.729 ± 0.199 | 6.279 ± 0.081 | 4.305 ± 0.146 | 0.00% ± 0.00 |
| Prompt / seed | Merged base · no new LoRA | Rank32 × 0.5 | Rank32 × 0.75 | Rank32 × 1.0 |
|---|---|---|---|---|
| Seed 1701 The door is still open. I put your coat on the chair. Tell me when you are ready. Exact GENERAL / SCRIPT promptGENERAL: One adult speaker in a clean recording. Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] The door is still open. (Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] I put your coat on the chair. (Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Tell me when you are ready. | Target 0.490 · CE 6.25 · Genuine 4.32 · WER 0.0% | Target 0.626 · CE 6.20 · Genuine 3.79 · WER 0.0% | Target 0.653 · CE 6.29 · Genuine 4.02 · WER 0.0% | Target 0.669 · CE 6.27 · Genuine 4.33 · WER 0.0% |
| Seed 1702 The door is still open. I put your coat on the chair. Tell me when you are ready. Exact GENERAL / SCRIPT promptGENERAL: One adult speaker in a clean recording. Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] The door is still open. (Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] I put your coat on the chair. (Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Tell me when you are ready. | Target 0.491 · CE 6.12 · Genuine 4.01 · WER 0.0% | Target 0.418 · CE 6.17 · Genuine 4.28 · WER 0.0% | Target 0.496 · CE 6.22 · Genuine 4.08 · WER 0.0% | Target 0.475 · CE 6.21 · Genuine 4.11 · WER 0.0% |
| Seed 1701 I found the little box. It was beside the old window. We can look inside together. Exact GENERAL / SCRIPT promptGENERAL: One adult speaker in a clean recording. Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] I found the little box. (Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] It was beside the old window. (Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] We can look inside together. | Target 0.780 · CE 6.39 · Genuine 4.18 · WER 0.0% | Target 0.775 · CE 6.38 · Genuine 4.39 · WER 0.0% | Target 0.839 · CE 6.36 · Genuine 4.28 · WER 0.0% | Target 0.856 · CE 6.40 · Genuine 4.33 · WER 0.0% |
| Seed 1702 I found the little box. It was beside the old window. We can look inside together. Exact GENERAL / SCRIPT promptGENERAL: One adult speaker in a clean recording. Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] I found the little box. (Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] It was beside the old window. (Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] We can look inside together. | Target 0.935 · CE 6.25 · Genuine 4.34 · WER 0.0% | Target 0.805 · CE 6.32 · Genuine 4.34 · WER 0.0% | Target 0.734 · CE 6.47 · Genuine 4.49 · WER 0.0% | Target 0.917 · CE 6.24 · Genuine 4.46 · WER 0.0% |
Fixed speaker reference:
| Condition | n | Target score ↑ | Content Enjoyment ↑ | Genuineness ↑ | WER ↓ |
|---|---|---|---|---|---|
| Merged base, no bucket adapter | 4 | 0.623 ± 0.100 | 6.299 ± 0.107 | 4.314 ± 0.166 | 0.00% ± 0.00 |
| Rank32 × 0.5 | 4 | 0.569 ± 0.091 | 6.326 ± 0.087 | 4.310 ± 0.109 | 0.00% ± 0.00 |
| Rank32 × 0.75 | 4 | 0.582 ± 0.137 | 6.328 ± 0.071 | 4.498 ± 0.178 | 0.00% ± 0.00 |
| Rank32 × 1.0 | 4 | 0.534 ± 0.139 | 6.255 ± 0.150 | 4.361 ± 0.108 | 0.00% ± 0.00 |
| Prompt / seed | Merged base · no new LoRA | Rank32 × 0.5 | Rank32 × 0.75 | Rank32 × 1.0 |
|---|---|---|---|---|
| Seed 1701 Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist. Exact GENERAL / SCRIPT promptGENERAL: One adult speaker in a clean recording. Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Die Tür ist noch offen. (Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt. (Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Sag mir, wenn du bereit bist. | Target 0.478 · CE 6.14 · Genuine 4.24 · WER 0.0% | Target 0.539 · CE 6.36 · Genuine 4.41 · WER 0.0% | Target 0.537 · CE 6.36 · Genuine 4.51 · WER 0.0% | Target 0.452 · CE 6.12 · Genuine 4.22 · WER 0.0% |
| Seed 1702 Die Tür ist noch offen. Ich habe deinen Mantel auf den Stuhl gelegt. Sag mir, wenn du bereit bist. Exact GENERAL / SCRIPT promptGENERAL: One adult speaker in a clean recording. Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Die Tür ist noch offen. (Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Ich habe deinen Mantel auf den Stuhl gelegt. (Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Sag mir, wenn du bereit bist. | Target 0.650 · CE 6.32 · Genuine 4.53 · WER 0.0% | Target 0.472 · CE 6.20 · Genuine 4.39 · WER 0.0% | Target 0.410 · CE 6.32 · Genuine 4.62 · WER 0.0% | Target 0.398 · CE 6.13 · Genuine 4.37 · WER 0.0% |
| Seed 1701 Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen. Exact GENERAL / SCRIPT promptGENERAL: One adult speaker in a clean recording. Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden. (Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Sie stand neben dem alten Fenster. (Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Wir können gemeinsam hineinschauen. | Target 0.706 · CE 6.37 · Genuine 4.14 · WER 0.0% | Target 0.574 · CE 6.36 · Genuine 4.22 · WER 0.0% | Target 0.718 · CE 6.40 · Genuine 4.24 · WER 0.0% | Target 0.710 · CE 6.36 · Genuine 4.48 · WER 0.0% |
| Seed 1702 Ich habe die kleine Schachtel gefunden. Sie stand neben dem alten Fenster. Wir können gemeinsam hineinschauen. Exact GENERAL / SCRIPT promptGENERAL: One adult speaker in a clean recording. Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Ich habe die kleine Schachtel gefunden. (Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Sie stand neben dem alten Fenster. (Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Wir können gemeinsam hineinschauen. | Target 0.656 · CE 6.36 · Genuine 4.34 · WER 0.0% | Target 0.690 · CE 6.39 · Genuine 4.21 · WER 0.0% | Target 0.662 · CE 6.24 · Genuine 4.62 · WER 0.0% | Target 0.574 · CE 6.40 · Genuine 4.38 · WER 0.0% |
Fixed speaker reference:
| Condition | n | Target score ↑ | Content Enjoyment ↑ | Genuineness ↑ | WER ↓ |
|---|---|---|---|---|---|
| Merged base, no bucket adapter | 4 | 0.485 ± 0.082 | 6.106 ± 0.155 | 4.523 ± 0.241 | 0.00% ± 0.00 |
| Rank32 × 0.5 | 4 | 0.481 ± 0.044 | 6.205 ± 0.121 | 4.523 ± 0.243 | 0.00% ± 0.00 |
| Rank32 × 0.75 | 4 | 0.440 ± 0.056 | 6.116 ± 0.084 | 4.313 ± 0.192 | 2.63% ± 5.26 |
| Rank32 × 1.0 | 4 | 0.467 ± 0.069 | 6.122 ± 0.052 | 4.382 ± 0.135 | 0.00% ± 0.00 |
| Prompt / seed | Merged base · no new LoRA | Rank32 × 0.5 | Rank32 × 0.75 | Rank32 × 1.0 |
|---|---|---|---|---|
| Seed 1701 La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt. Exact GENERAL / SCRIPT promptGENERAL: One adult speaker in a clean recording. Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La porte est encore ouverte. (Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise. (Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Dis-moi quand tu es prêt. | Target 0.388 · CE 6.17 · Genuine 4.52 · WER 0.0% | Target 0.469 · CE 6.16 · Genuine 4.49 · WER 0.0% | Target 0.370 · CE 6.12 · Genuine 4.04 · WER 0.0% | Target 0.536 · CE 6.06 · Genuine 4.31 · WER 0.0% |
| Seed 1702 La porte est encore ouverte. J’ai posé ton manteau sur la chaise. Dis-moi quand tu es prêt. Exact GENERAL / SCRIPT promptGENERAL: One adult speaker in a clean recording. Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La porte est encore ouverte. (Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] J’ai posé ton manteau sur la chaise. (Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Dis-moi quand tu es prêt. | Target 0.446 · CE 5.97 · Genuine 4.23 · WER 0.0% | Target 0.431 · CE 6.10 · Genuine 4.22 · WER 0.0% | Target 0.423 · CE 6.01 · Genuine 4.33 · WER 0.0% | Target 0.389 · CE 6.13 · Genuine 4.30 · WER 0.0% |
| Seed 1701 J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient. Exact GENERAL / SCRIPT promptGENERAL: One adult speaker in a clean recording. Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] J’ai trouvé la petite boîte. (Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Elle était près de la vieille fenêtre. (Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient. | Target 0.557 · CE 6.29 · Genuine 4.52 · WER 0.0% | Target 0.536 · CE 6.18 · Genuine 4.57 · WER 0.0% | Target 0.496 · CE 6.11 · Genuine 4.44 · WER 10.5% | Target 0.432 · CE 6.19 · Genuine 4.58 · WER 0.0% |
| Seed 1702 J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient. Exact GENERAL / SCRIPT promptGENERAL: One adult speaker in a clean recording. Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] J’ai trouvé la petite boîte. (Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Elle était près de la vieille fenêtre. (Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Nous pouvons regarder ensemble ce qu’elle contient. | Target 0.550 · CE 5.98 · Genuine 4.82 · WER 0.0% | Target 0.488 · CE 6.38 · Genuine 4.81 · WER 0.0% | Target 0.472 · CE 6.22 · Genuine 4.44 · WER 0.0% | Target 0.512 · CE 6.11 · Genuine 4.33 · WER 0.0% |
Fixed speaker reference:
| Condition | n | Target score ↑ | Content Enjoyment ↑ | Genuineness ↑ | WER ↓ |
|---|---|---|---|---|---|
| Merged base, no bucket adapter | 4 | 0.595 ± 0.166 | 6.160 ± 0.076 | 4.339 ± 0.058 | 0.00% ± 0.00 |
| Rank32 × 0.5 | 4 | 0.612 ± 0.097 | 6.128 ± 0.058 | 4.258 ± 0.084 | 0.00% ± 0.00 |
| Rank32 × 0.75 | 4 | 0.620 ± 0.243 | 6.094 ± 0.044 | 4.299 ± 0.049 | 0.00% ± 0.00 |
| Rank32 × 1.0 | 4 | 0.603 ± 0.177 | 6.094 ± 0.096 | 4.319 ± 0.244 | 0.00% ± 0.00 |
| Prompt / seed | Merged base · no new LoRA | Rank32 × 0.5 | Rank32 × 0.75 | Rank32 × 1.0 |
|---|---|---|---|---|
| Seed 1701 De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent. Exact GENERAL / SCRIPT promptGENERAL: One adult speaker in a clean recording. Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] De deur staat nog open. (Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd. (Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Zeg maar wanneer je klaar bent. | Target 0.759 · CE 6.25 · Genuine 4.31 · WER 0.0% | Target 0.474 · CE 6.13 · Genuine 4.30 · WER 0.0% | Target 0.418 · CE 6.03 · Genuine 4.29 · WER 0.0% | Target 0.467 · CE 6.04 · Genuine 4.58 · WER 0.0% |
| Seed 1702 De deur staat nog open. Ik heb je jas op de stoel gelegd. Zeg maar wanneer je klaar bent. Exact GENERAL / SCRIPT promptGENERAL: One adult speaker in a clean recording. Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] De deur staat nog open. (Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Ik heb je jas op de stoel gelegd. (Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Zeg maar wanneer je klaar bent. | Target 0.366 · CE 6.17 · Genuine 4.41 · WER 0.0% | Target 0.621 · CE 6.07 · Genuine 4.24 · WER 0.0% | Target 0.402 · CE 6.12 · Genuine 4.26 · WER 0.0% | Target 0.495 · CE 6.05 · Genuine 4.22 · WER 0.0% |
| Seed 1701 Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken. Exact GENERAL / SCRIPT promptGENERAL: One adult speaker in a clean recording. Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Ik heb het kleine doosje gevonden. (Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Het stond naast het oude raam. (Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] We kunnen er samen in kijken. | Target 0.646 · CE 6.14 · Genuine 4.28 · WER 0.0% | Target 0.698 · CE 6.10 · Genuine 4.15 · WER 0.0% | Target 0.834 · CE 6.10 · Genuine 4.27 · WER 0.0% | Target 0.856 · CE 6.05 · Genuine 4.03 · WER 0.0% |
| Seed 1702 Ik heb het kleine doosje gevonden. Het stond naast het oude raam. We kunnen er samen in kijken. Exact GENERAL / SCRIPT promptGENERAL: One adult speaker in a clean recording. Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Ik heb het kleine doosje gevonden. (Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Het stond naast het oude raam. (Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] We kunnen er samen in kijken. | Target 0.608 · CE 6.07 · Genuine 4.36 · WER 0.0% | Target 0.655 · CE 6.21 · Genuine 4.34 · WER 0.0% | Target 0.826 · CE 6.13 · Genuine 4.37 · WER 0.0% | Target 0.592 · CE 6.24 · Genuine 4.44 · WER 0.0% |
Fixed speaker reference:
| Condition | n | Target score ↑ | Content Enjoyment ↑ | Genuineness ↑ | WER ↓ |
|---|---|---|---|---|---|
| Merged base, no bucket adapter | 4 | 0.538 ± 0.085 | 6.046 ± 0.044 | 4.308 ± 0.265 | 1.47% ± 2.94 |
| Rank32 × 0.5 | 4 | 0.428 ± 0.064 | 6.106 ± 0.062 | 4.319 ± 0.239 | 1.47% ± 2.94 |
| Rank32 × 0.75 | 4 | 0.452 ± 0.083 | 5.938 ± 0.355 | 4.310 ± 0.394 | 0.00% ± 0.00 |
| Rank32 × 1.0 | 4 | 0.531 ± 0.169 | 6.164 ± 0.105 | 4.435 ± 0.320 | 0.00% ± 0.00 |
| Prompt / seed | Merged base · no new LoRA | Rank32 × 0.5 | Rank32 × 0.75 | Rank32 × 1.0 |
|---|---|---|---|---|
| Seed 1701 La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo. Exact GENERAL / SCRIPT promptGENERAL: One adult speaker in a clean recording. Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La puerta sigue abierta. (Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] He dejado tu abrigo sobre la silla. (Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Dime cuándo estás listo. | Target 0.567 · CE 6.09 · Genuine 4.30 · WER 0.0% | Target 0.434 · CE 6.12 · Genuine 4.06 · WER 0.0% | Target 0.377 · CE 5.42 · Genuine 4.34 · WER 0.0% | Target 0.379 · CE 6.05 · Genuine 4.03 · WER 0.0% |
| Seed 1702 La puerta sigue abierta. He dejado tu abrigo sobre la silla. Dime cuándo estás listo. Exact GENERAL / SCRIPT promptGENERAL: One adult speaker in a clean recording. Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La puerta sigue abierta. (Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] He dejado tu abrigo sobre la silla. (Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Dime cuándo estás listo. | Target 0.465 · CE 6.03 · Genuine 3.94 · WER 0.0% | Target 0.349 · CE 6.01 · Genuine 4.26 · WER 0.0% | Target 0.393 · CE 6.11 · Genuine 3.74 · WER 0.0% | Target 0.426 · CE 6.09 · Genuine 4.38 · WER 0.0% |
| Seed 1701 He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro. Exact GENERAL / SCRIPT promptGENERAL: One adult speaker in a clean recording. Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] He encontrado la cajita. (Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Estaba junto a la ventana vieja. (Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro. | Target 0.645 · CE 5.99 · Genuine 4.43 · WER 5.9% | Target 0.420 · CE 6.14 · Genuine 4.33 · WER 0.0% | Target 0.555 · CE 6.00 · Genuine 4.60 · WER 0.0% | Target 0.563 · CE 6.24 · Genuine 4.53 · WER 0.0% |
| Seed 1702 He encontrado la cajita. Estaba junto a la ventana vieja. Podemos mirar juntos lo que hay dentro. Exact GENERAL / SCRIPT promptGENERAL: One adult speaker in a clean recording. Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] He encontrado la cajita. (Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.8 seconds duration] Estaba junto a la ventana vieja. (Intense Amusement. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Podemos mirar juntos lo que hay dentro. | Target 0.473 · CE 6.08 · Genuine 4.56 · WER 0.0% | Target 0.507 · CE 6.15 · Genuine 4.63 · WER 5.9% | Target 0.482 · CE 6.22 · Genuine 4.56 · WER 0.0% | Target 0.755 · CE 6.27 · Genuine 4.80 · WER 0.0% |