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: Held-out target improves against matched CFG and best manual-only configuration; QA passes. Bucket scale 0.75, CFG 4.0.
| Held-out condition | n | Target z↑ | Raw head | CE↑ | Genuine↑ | Voice↑ | WER↓ | Valid | QA |
|---|---|---|---|---|---|---|---|---|---|
| standard | 50 | -0.344 | 0.059 | 6.221 | 4.423 | 0.594 | 0.010 | 1.000 | pass |
| manual | 50 | 0.150 | 0.343 | 6.027 | 4.489 | 0.602 | 0.016 | 1.000 | pass |
| adapter | 50 | 0.890 | 0.769 | 5.799 | 4.239 | 0.550 | 0.047 | 0.980 | pass |
| matched_no_adapter | 50 | 0.527 | 0.560 | 5.922 | 4.388 | 0.591 | 0.016 | 1.000 | pass |
Incremental adapter gain versus same-CFG control: 0.363 z;95% cluster interval [0.21553710562046097, 0.5911914129008978]. Versus best manual-only: 0.741 z. Manual-only versus standard: 0.494 z.
| Scale | CFG | Target z | CE | Genuine | WER | Provisional QA |
|---|---|---|---|---|---|---|
| 0.0 | 1.0 | -0.327 | 6.164 | 4.560 | 0.000 | True |
| 0.5 | 1.0 | -0.308 | 6.136 | 4.505 | 0.006 | True |
| 0.75 | 1.0 | -0.314 | 6.135 | 4.508 | 0.003 | True |
| 1.0 | 1.0 | -0.344 | 6.158 | 4.512 | 0.003 | True |
| 0.0 | 1.25 | -0.272 | 6.094 | 4.581 | 0.011 | True |
| 0.5 | 1.25 | -0.252 | 6.150 | 4.539 | 0.000 | True |
| 0.75 | 1.25 | -0.240 | 6.079 | 4.552 | 0.011 | True |
| 1.0 | 1.25 | -0.220 | 6.069 | 4.584 | 0.007 | True |
| 0.0 | 1.5 | -0.267 | 6.027 | 4.556 | 0.000 | True |
| 0.5 | 1.5 | -0.138 | 6.072 | 4.521 | 0.008 | True |
| 0.75 | 1.5 | -0.177 | 6.056 | 4.567 | 0.009 | True |
| 1.0 | 1.5 | -0.051 | 6.032 | 4.587 | 0.000 | True |
| 0.0 | 2.0 | -0.138 | 6.089 | 4.613 | 0.000 | True |
| 0.5 | 2.0 | -0.058 | 5.955 | 4.562 | 0.011 | True |
| 0.75 | 2.0 | 0.231 | 5.972 | 4.566 | 0.006 | True |
| 1.0 | 2.0 | 0.116 | 5.968 | 4.558 | 0.014 | True |
| 0.0 | 2.5 | 0.122 | 6.029 | 4.602 | 0.014 | True |
| 0.5 | 2.5 | 0.060 | 5.986 | 4.582 | 0.023 | True |
| 0.75 | 2.5 | 0.390 | 5.888 | 4.533 | 0.003 | True |
| 1.0 | 2.5 | 0.270 | 5.885 | 4.472 | 0.025 | True |
| 0.0 | 3.0 | 0.392 | 5.944 | 4.575 | 0.006 | True |
| 0.5 | 3.0 | 0.410 | 5.890 | 4.565 | 0.017 | True |
| 0.75 | 3.0 | 0.472 | 5.840 | 4.484 | 0.019 | True |
| 1.0 | 3.0 | 0.641 | 5.784 | 4.484 | 0.003 | True |
| 0.0 | 4.0 | 1.321 | 5.841 | 4.402 | 0.006 | False |
| 0.5 | 4.0 | 1.416 | 5.755 | 4.326 | 0.026 | False |
| 0.75 | 4.0 | 0.985 | 5.736 | 4.451 | 0.014 | True |
| 1.0 | 4.0 | 1.315 | 5.667 | 4.360 | 0.009 | False |
| 0.0 | 5.0 | 1.328 | 5.789 | 4.384 | 0.017 | False |
| 0.5 | 5.0 | 1.671 | 5.710 | 4.347 | 0.039 | False |
| 0.75 | 5.0 | 1.535 | 5.624 | 4.282 | 0.038 | False |
| 1.0 | 5.0 | 1.527 | 5.595 | 4.333 | 0.015 | False |
direct_manual_recipe
{
"id": "manual-abbf6e2776b7",
"general": "The performance must convey a sense of desperate, near-breaking vulnerability. The pain is less about the injury and more about the inability to control the reaction to it.",
"inline": [
"Begin with an extremely fragile, almost breathy onset, as if the speaker cannot summon the strength to project the initial sound.",
"Introduce brief, uncontrolled vocal fry or vocal breaks mid-sentence, punctuating the speech with involuntary physical reactions.",
"Conclude with a forced, strained projection on the final word, a momentary surge of desperate will overriding the physical distress."
],
"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 Pain. 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": 43, "de": 41, "en": 55, "nl": 6, "es": 46}. 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 | 0.369 ± 0.017 | 6.170 ± 0.135 | 4.456 ± 0.173 | 0.59% ± 2.63 |
| Rank32 × 0.5 | 20 | 0.370 ± 0.014 | 6.149 ± 0.173 | 4.410 ± 0.264 | 0.00% ± 0.00 |
| Rank32 × 0.75 | 20 | 0.373 ± 0.019 | 6.179 ± 0.128 | 4.454 ± 0.177 | 0.79% ± 2.58 |
| Rank32 × 1.0 | 20 | 0.367 ± 0.015 | 6.152 ± 0.107 | 4.354 ± 0.260 | 0.00% ± 0.00 |
Fixed speaker reference:
| Condition | n | Target score ↑ | Content Enjoyment ↑ | Genuineness ↑ | WER ↓ |
|---|---|---|---|---|---|
| Merged base, no bucket adapter | 4 | 0.388 ± 0.008 | 6.281 ± 0.136 | 4.344 ± 0.321 | 0.00% ± 0.00 |
| Rank32 × 0.5 | 4 | 0.382 ± 0.010 | 6.322 ± 0.076 | 4.398 ± 0.352 | 0.00% ± 0.00 |
| Rank32 × 0.75 | 4 | 0.396 ± 0.013 | 6.279 ± 0.120 | 4.404 ± 0.266 | 0.00% ± 0.00 |
| Rank32 × 1.0 | 4 | 0.389 ± 0.004 | 6.211 ± 0.067 | 4.292 ± 0.331 | 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 Pain. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Pain. 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 Pain. 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 Pain. 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 0.383 · CE 6.36 · Genuine 4.19 · WER 0.0% | Target 0.384 · CE 6.25 · Genuine 4.11 · WER 0.0% | Target 0.405 · CE 6.25 · Genuine 4.43 · WER 0.0% | Target 0.388 · CE 6.18 · Genuine 3.89 · 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 Pain. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Pain. 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 Pain. 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 Pain. 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 0.387 · CE 6.08 · Genuine 3.97 · WER 0.0% | Target 0.376 · CE 6.28 · Genuine 4.08 · WER 0.0% | Target 0.392 · CE 6.17 · Genuine 4.04 · WER 0.0% | Target 0.384 · CE 6.20 · Genuine 4.19 · 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 Pain. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Pain. 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 Pain. 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 Pain. 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 0.381 · CE 6.35 · Genuine 4.66 · WER 0.0% | Target 0.395 · CE 6.36 · Genuine 4.72 · WER 0.0% | Target 0.408 · CE 6.45 · Genuine 4.68 · WER 0.0% | Target 0.393 · CE 6.16 · Genuine 4.43 · 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 Pain. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Pain. 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 Pain. 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 Pain. 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 0.399 · CE 6.33 · Genuine 4.56 · WER 0.0% | Target 0.372 · CE 6.41 · Genuine 4.68 · WER 0.0% | Target 0.380 · CE 6.25 · Genuine 4.46 · WER 0.0% | Target 0.389 · CE 6.31 · Genuine 4.66 · WER 0.0% |
Fixed speaker reference:
| Condition | n | Target score ↑ | Content Enjoyment ↑ | Genuineness ↑ | WER ↓ |
|---|---|---|---|---|---|
| Merged base, no bucket adapter | 4 | 0.378 ± 0.011 | 6.273 ± 0.085 | 4.601 ± 0.090 | 0.00% ± 0.00 |
| Rank32 × 0.5 | 4 | 0.373 ± 0.012 | 6.245 ± 0.119 | 4.536 ± 0.086 | 0.00% ± 0.00 |
| Rank32 × 0.75 | 4 | 0.371 ± 0.018 | 6.283 ± 0.141 | 4.490 ± 0.149 | 2.63% ± 5.26 |
| Rank32 × 1.0 | 4 | 0.368 ± 0.014 | 6.267 ± 0.113 | 4.515 ± 0.141 | 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 Pain. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Pain. 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 Pain. 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 Pain. 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 0.380 · CE 6.17 · Genuine 4.48 · WER 0.0% | Target 0.388 · CE 6.18 · Genuine 4.46 · WER 0.0% | Target 0.385 · CE 6.10 · Genuine 4.46 · WER 0.0% | Target 0.384 · CE 6.11 · Genuine 4.33 · 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 Pain. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Pain. 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 Pain. 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 Pain. 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 0.392 · CE 6.26 · Genuine 4.58 · WER 0.0% | Target 0.361 · CE 6.11 · Genuine 4.52 · WER 0.0% | Target 0.387 · CE 6.27 · Genuine 4.71 · WER 10.5% | Target 0.376 · CE 6.26 · Genuine 4.65 · 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 Pain. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Pain. 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 Pain. 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 Pain. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Wir können gemeinsam hineinschauen. | Target 0.374 · CE 6.38 · Genuine 4.67 · WER 0.0% | Target 0.376 · CE 6.33 · Genuine 4.51 · WER 0.0% | Target 0.351 · CE 6.44 · Genuine 4.38 · WER 0.0% | Target 0.355 · CE 6.33 · Genuine 4.49 · 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 Pain. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Pain. 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 Pain. 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 Pain. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Wir können gemeinsam hineinschauen. | Target 0.365 · CE 6.28 · Genuine 4.67 · WER 0.0% | Target 0.367 · CE 6.36 · Genuine 4.66 · WER 0.0% | Target 0.363 · CE 6.32 · Genuine 4.42 · WER 0.0% | Target 0.357 · CE 6.36 · Genuine 4.60 · WER 0.0% |
Fixed speaker reference:
| Condition | n | Target score ↑ | Content Enjoyment ↑ | Genuineness ↑ | WER ↓ |
|---|---|---|---|---|---|
| Merged base, no bucket adapter | 4 | 0.360 ± 0.013 | 6.124 ± 0.041 | 4.424 ± 0.078 | 0.00% ± 0.00 |
| Rank32 × 0.5 | 4 | 0.365 ± 0.014 | 6.033 ± 0.102 | 4.462 ± 0.128 | 0.00% ± 0.00 |
| Rank32 × 0.75 | 4 | 0.375 ± 0.018 | 6.170 ± 0.063 | 4.586 ± 0.151 | 1.32% ± 2.63 |
| Rank32 × 1.0 | 4 | 0.367 ± 0.002 | 6.078 ± 0.030 | 4.451 ± 0.248 | 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 Pain. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Pain. 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 Pain. 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 Pain. 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 0.360 · CE 6.15 · Genuine 4.43 · WER 0.0% | Target 0.370 · CE 5.88 · Genuine 4.28 · WER 0.0% | Target 0.369 · CE 6.14 · Genuine 4.66 · WER 0.0% | Target 0.363 · CE 6.07 · Genuine 4.58 · 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 Pain. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Pain. 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 Pain. 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 Pain. 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 0.342 · CE 6.06 · Genuine 4.51 · WER 0.0% | Target 0.346 · CE 6.09 · Genuine 4.48 · WER 0.0% | Target 0.357 · CE 6.12 · Genuine 4.42 · WER 0.0% | Target 0.367 · CE 6.05 · Genuine 4.08 · 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 Pain. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Pain. 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 Pain. 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 Pain. 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 0.373 · CE 6.14 · Genuine 4.44 · WER 0.0% | Target 0.377 · CE 6.08 · Genuine 4.56 · WER 0.0% | Target 0.399 · CE 6.26 · Genuine 4.76 · WER 5.3% | Target 0.368 · CE 6.08 · Genuine 4.60 · 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 Pain. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Pain. 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 Pain. 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 Pain. 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 0.365 · CE 6.14 · Genuine 4.32 · WER 0.0% | Target 0.368 · CE 6.09 · Genuine 4.53 · WER 0.0% | Target 0.374 · CE 6.16 · Genuine 4.51 · WER 0.0% | Target 0.368 · CE 6.12 · Genuine 4.55 · WER 0.0% |
Fixed speaker reference:
| Condition | n | Target score ↑ | Content Enjoyment ↑ | Genuineness ↑ | WER ↓ |
|---|---|---|---|---|---|
| Merged base, no bucket adapter | 4 | 0.351 ± 0.004 | 6.187 ± 0.034 | 4.418 ± 0.091 | 0.00% ± 0.00 |
| Rank32 × 0.5 | 4 | 0.367 ± 0.019 | 6.108 ± 0.118 | 4.328 ± 0.099 | 0.00% ± 0.00 |
| Rank32 × 0.75 | 4 | 0.358 ± 0.012 | 6.056 ± 0.068 | 4.410 ± 0.093 | 0.00% ± 0.00 |
| Rank32 × 1.0 | 4 | 0.356 ± 0.010 | 6.156 ± 0.059 | 4.281 ± 0.146 | 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 Pain. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Pain. 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 Pain. 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 Pain. 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 0.347 · CE 6.15 · Genuine 4.36 · WER 0.0% | Target 0.356 · CE 6.17 · Genuine 4.18 · WER 0.0% | Target 0.371 · CE 6.01 · Genuine 4.53 · WER 0.0% | Target 0.353 · CE 6.08 · Genuine 4.08 · 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 Pain. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Pain. 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 Pain. 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 Pain. 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 0.347 · CE 6.17 · Genuine 4.32 · WER 0.0% | Target 0.388 · CE 6.13 · Genuine 4.37 · WER 0.0% | Target 0.366 · CE 6.13 · Genuine 4.39 · WER 0.0% | Target 0.348 · CE 6.14 · Genuine 4.27 · 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 Pain. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Pain. 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 Pain. 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 Pain. 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 0.354 · CE 6.23 · Genuine 4.51 · WER 0.0% | Target 0.376 · CE 6.19 · Genuine 4.38 · WER 0.0% | Target 0.345 · CE 5.99 · Genuine 4.31 · WER 0.0% | Target 0.370 · CE 6.20 · Genuine 4.41 · 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 Pain. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Pain. 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 Pain. 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 Pain. 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 0.355 · CE 6.20 · Genuine 4.48 · WER 0.0% | Target 0.347 · CE 5.94 · Genuine 4.38 · WER 0.0% | Target 0.351 · CE 6.10 · Genuine 4.41 · WER 0.0% | Target 0.351 · CE 6.21 · Genuine 4.37 · WER 0.0% |
Fixed speaker reference:
| Condition | n | Target score ↑ | Content Enjoyment ↑ | Genuineness ↑ | WER ↓ |
|---|---|---|---|---|---|
| Merged base, no bucket adapter | 4 | 0.369 ± 0.015 | 5.985 ± 0.086 | 4.493 ± 0.119 | 2.94% ± 5.88 |
| Rank32 × 0.5 | 4 | 0.362 ± 0.010 | 6.036 ± 0.237 | 4.328 ± 0.491 | 0.00% ± 0.00 |
| Rank32 × 0.75 | 4 | 0.367 ± 0.020 | 6.106 ± 0.074 | 4.382 ± 0.188 | 0.00% ± 0.00 |
| Rank32 × 1.0 | 4 | 0.357 ± 0.012 | 6.050 ± 0.084 | 4.228 ± 0.367 | 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 Pain. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Pain. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La puerta sigue abierta. (Intense Pain. 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 Pain. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Dime cuándo estás listo. | Target 0.388 · CE 5.90 · Genuine 4.60 · WER 0.0% | Target 0.357 · CE 5.94 · Genuine 3.89 · WER 0.0% | Target 0.395 · CE 6.16 · Genuine 4.38 · WER 0.0% | Target 0.358 · CE 5.94 · Genuine 4.01 · 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 Pain. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Pain. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La puerta sigue abierta. (Intense Pain. 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 Pain. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Dime cuándo estás listo. | Target 0.369 · CE 5.92 · Genuine 4.33 · WER 0.0% | Target 0.357 · CE 5.80 · Genuine 3.92 · WER 0.0% | Target 0.365 · CE 6.05 · Genuine 4.12 · WER 0.0% | Target 0.351 · CE 6.08 · Genuine 3.84 · 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 Pain. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Pain. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] He encontrado la cajita. (Intense Pain. 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 Pain. 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 0.367 · CE 6.05 · Genuine 4.56 · WER 11.8% | Target 0.377 · CE 6.36 · Genuine 4.80 · WER 0.0% | Target 0.347 · CE 6.18 · Genuine 4.48 · WER 0.0% | Target 0.373 · CE 6.13 · Genuine 4.63 · 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 Pain. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Pain. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] He encontrado la cajita. (Intense Pain. 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 Pain. 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 0.350 · CE 6.07 · Genuine 4.49 · WER 0.0% | Target 0.357 · CE 6.05 · Genuine 4.71 · WER 0.0% | Target 0.361 · CE 6.04 · Genuine 4.55 · WER 0.0% | Target 0.347 · CE 6.05 · Genuine 4.43 · WER 0.0% |