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: Use manual-only configuration; no clear qualifying incremental adapter gain. Bucket scale 0.0, CFG 3.0.
| Held-out condition | n | Target z↑ | Raw head | CE↑ | Genuine↑ | Voice↑ | WER↓ | Valid | QA |
|---|---|---|---|---|---|---|---|---|---|
| standard | 50 | -0.379 | 0.065 | 6.275 | 4.343 | 0.556 | 0.006 | 1.000 | pass |
| manual | 50 | -0.119 | 0.171 | 6.068 | 4.535 | 0.551 | 0.021 | 1.000 | pass |
| adapter | 50 | -0.088 | 0.184 | 5.979 | 4.481 | 0.510 | 0.018 | 0.980 | pass |
| matched_no_adapter | 50 | -0.119 | 0.171 | 6.068 | 4.535 | 0.551 | 0.021 | 1.000 | pass |
Incremental adapter gain versus same-CFG control: 0.031 z;95% cluster interval [-0.03822265684648564, 0.11277343922508798]. Versus best manual-only: 0.031 z. Manual-only versus standard: 0.260 z.
| Scale | CFG | Target z | CE | Genuine | WER | Provisional QA |
|---|---|---|---|---|---|---|
| 0.0 | 1.0 | -0.385 | 6.139 | 4.422 | 0.005 | True |
| 0.5 | 1.0 | -0.367 | 6.130 | 4.409 | 0.006 | True |
| 0.75 | 1.0 | -0.388 | 6.151 | 4.449 | 0.006 | True |
| 1.0 | 1.0 | -0.360 | 6.102 | 4.434 | 0.000 | True |
| 0.0 | 1.25 | -0.348 | 6.112 | 4.422 | 0.003 | True |
| 0.5 | 1.25 | -0.354 | 6.100 | 4.485 | 0.016 | True |
| 0.75 | 1.25 | -0.359 | 6.104 | 4.412 | 0.003 | True |
| 1.0 | 1.25 | -0.351 | 6.100 | 4.447 | 0.003 | True |
| 0.0 | 1.5 | -0.332 | 6.116 | 4.558 | 0.008 | True |
| 0.5 | 1.5 | -0.305 | 6.116 | 4.530 | 0.009 | True |
| 0.75 | 1.5 | -0.329 | 6.158 | 4.540 | 0.000 | True |
| 1.0 | 1.5 | -0.317 | 6.078 | 4.466 | 0.006 | True |
| 0.0 | 2.0 | -0.257 | 6.005 | 4.583 | 0.000 | True |
| 0.5 | 2.0 | -0.211 | 6.085 | 4.550 | 0.000 | True |
| 0.75 | 2.0 | -0.248 | 6.065 | 4.528 | 0.000 | True |
| 1.0 | 2.0 | -0.262 | 6.065 | 4.566 | 0.006 | True |
| 0.0 | 2.5 | -0.230 | 5.973 | 4.603 | 0.003 | True |
| 0.5 | 2.5 | -0.240 | 6.019 | 4.567 | 0.014 | False |
| 0.75 | 2.5 | -0.227 | 5.956 | 4.546 | 0.015 | True |
| 1.0 | 2.5 | -0.187 | 5.910 | 4.584 | 0.009 | True |
| 0.0 | 3.0 | -0.184 | 5.919 | 4.565 | 0.000 | True |
| 0.5 | 3.0 | -0.217 | 5.987 | 4.532 | 0.006 | True |
| 0.75 | 3.0 | -0.185 | 5.859 | 4.510 | 0.023 | True |
| 1.0 | 3.0 | -0.229 | 5.810 | 4.501 | 0.000 | True |
| 0.0 | 4.0 | -0.021 | 5.872 | 4.445 | 0.016 | False |
| 0.5 | 4.0 | -0.139 | 5.670 | 4.416 | 0.034 | False |
| 0.75 | 4.0 | -0.084 | 5.686 | 4.430 | 0.017 | False |
| 1.0 | 4.0 | -0.131 | 5.674 | 4.394 | 0.016 | False |
| 0.0 | 5.0 | -0.047 | 5.665 | 4.377 | 0.033 | False |
| 0.5 | 5.0 | 0.278 | 5.709 | 4.387 | 0.033 | False |
| 0.75 | 5.0 | 0.125 | 5.637 | 4.379 | 0.009 | False |
| 1.0 | 5.0 | 0.012 | 5.548 | 4.288 | 0.076 | False |
direct_manual_recipe
{
"id": "manual-564f1f7c16dd",
"general": "Shift the energy from outright anger to profound, wounded disappointment; the bitterness should feel internal and brittle.",
"inline": [
"Begin with a slightly constricted throat sound, suggesting the words are physically difficult to produce.",
"Modulate the volume very slightly, a barely perceptible dip in projection that suggests withdrawal.",
"Let the final utterance carry a thin, high-frequency edge, like glass about to shatter under pressure."
],
"cfg": 3,
"recipe": "P05",
"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 Bitterness. 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: {"de": 46, "fr": 35, "es": 15, "en": 41, "nl": 3}. 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 | 0.353 ± 0.050 | 6.175 ± 0.123 | 4.404 ± 0.241 | 0.00% ± 0.00 |
| Rank32 × 0.5 | 20 | 0.340 ± 0.046 | 6.206 ± 0.112 | 4.337 ± 0.179 | 0.00% ± 0.00 |
| Rank32 × 0.75 | 20 | 0.352 ± 0.076 | 6.181 ± 0.144 | 4.330 ± 0.260 | 0.33% ± 1.49 |
| Rank32 × 1.0 | 20 | 0.345 ± 0.052 | 6.195 ± 0.138 | 4.295 ± 0.176 | 0.28% ± 1.24 |
Fixed speaker reference:
| Condition | n | Target score ↑ | Content Enjoyment ↑ | Genuineness ↑ | WER ↓ |
|---|---|---|---|---|---|
| Merged base, no bucket adapter | 4 | 0.303 ± 0.048 | 6.271 ± 0.107 | 4.281 ± 0.307 | 0.00% ± 0.00 |
| Rank32 × 0.5 | 4 | 0.297 ± 0.044 | 6.295 ± 0.108 | 4.183 ± 0.243 | 0.00% ± 0.00 |
| Rank32 × 0.75 | 4 | 0.309 ± 0.045 | 6.291 ± 0.117 | 4.103 ± 0.298 | 0.00% ± 0.00 |
| Rank32 × 1.0 | 4 | 0.311 ± 0.059 | 6.287 ± 0.143 | 4.267 ± 0.175 | 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 Bitterness. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Bitterness. 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 Bitterness. 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 Bitterness. 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 0.328 · CE 6.31 · Genuine 4.21 · WER 0.0% | Target 0.339 · CE 6.20 · Genuine 4.11 · WER 0.0% | Target 0.333 · CE 6.16 · Genuine 3.73 · WER 0.0% | Target 0.346 · CE 6.22 · Genuine 4.07 · 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 Bitterness. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Bitterness. 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 Bitterness. 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 Bitterness. 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 0.355 · CE 6.13 · Genuine 3.88 · WER 0.0% | Target 0.332 · CE 6.21 · Genuine 3.88 · WER 0.0% | Target 0.355 · CE 6.25 · Genuine 4.02 · WER 0.0% | Target 0.374 · CE 6.12 · Genuine 4.18 · 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 Bitterness. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Bitterness. 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 Bitterness. 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 Bitterness. 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 0.282 · CE 6.38 · Genuine 4.56 · WER 0.0% | Target 0.251 · CE 6.38 · Genuine 4.43 · WER 0.0% | Target 0.292 · CE 6.44 · Genuine 4.24 · WER 0.0% | Target 0.248 · CE 6.38 · Genuine 4.36 · 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 Bitterness. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Bitterness. 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 Bitterness. 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 Bitterness. 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 0.246 · CE 6.27 · Genuine 4.48 · WER 0.0% | Target 0.269 · CE 6.40 · Genuine 4.31 · WER 0.0% | Target 0.255 · CE 6.32 · Genuine 4.42 · WER 0.0% | Target 0.276 · CE 6.43 · Genuine 4.46 · WER 0.0% |
Fixed speaker reference:
| Condition | n | Target score ↑ | Content Enjoyment ↑ | Genuineness ↑ | WER ↓ |
|---|---|---|---|---|---|
| Merged base, no bucket adapter | 4 | 0.351 ± 0.039 | 6.303 ± 0.113 | 4.531 ± 0.109 | 0.00% ± 0.00 |
| Rank32 × 0.5 | 4 | 0.345 ± 0.029 | 6.291 ± 0.094 | 4.440 ± 0.141 | 0.00% ± 0.00 |
| Rank32 × 0.75 | 4 | 0.419 ± 0.139 | 6.330 ± 0.049 | 4.441 ± 0.192 | 0.00% ± 0.00 |
| Rank32 × 1.0 | 4 | 0.358 ± 0.060 | 6.235 ± 0.258 | 4.361 ± 0.179 | 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 Bitterness. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Bitterness. 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 Bitterness. 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 Bitterness. 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 0.376 · CE 6.19 · Genuine 4.61 · WER 0.0% | Target 0.352 · CE 6.28 · Genuine 4.29 · WER 0.0% | Target 0.620 · CE 6.36 · Genuine 4.21 · WER 0.0% | Target 0.392 · CE 6.09 · Genuine 4.52 · 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 Bitterness. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Bitterness. 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 Bitterness. 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 Bitterness. 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 6.22 · Genuine 4.53 · WER 0.0% | Target 0.371 · CE 6.24 · Genuine 4.63 · WER 0.0% | Target 0.402 · CE 6.27 · Genuine 4.67 · WER 0.0% | Target 0.424 · CE 5.94 · Genuine 4.46 · 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 Bitterness. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Bitterness. 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 Bitterness. 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 Bitterness. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Wir können gemeinsam hineinschauen. | Target 0.320 · CE 6.40 · Genuine 4.61 · WER 0.0% | Target 0.303 · CE 6.43 · Genuine 4.41 · WER 0.0% | Target 0.346 · CE 6.32 · Genuine 4.48 · WER 0.0% | Target 0.302 · CE 6.43 · Genuine 4.11 · 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 Bitterness. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Bitterness. 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 Bitterness. 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 Bitterness. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Wir können gemeinsam hineinschauen. | Target 0.317 · CE 6.40 · Genuine 4.38 · WER 0.0% | Target 0.354 · CE 6.22 · Genuine 4.42 · WER 0.0% | Target 0.310 · CE 6.38 · Genuine 4.41 · WER 0.0% | Target 0.313 · CE 6.48 · Genuine 4.36 · WER 0.0% |
Fixed speaker reference:
| Condition | n | Target score ↑ | Content Enjoyment ↑ | Genuineness ↑ | WER ↓ |
|---|---|---|---|---|---|
| Merged base, no bucket adapter | 4 | 0.375 ± 0.023 | 6.162 ± 0.025 | 4.649 ± 0.089 | 0.00% ± 0.00 |
| Rank32 × 0.5 | 4 | 0.333 ± 0.027 | 6.177 ± 0.035 | 4.360 ± 0.155 | 0.00% ± 0.00 |
| Rank32 × 0.75 | 4 | 0.332 ± 0.052 | 6.179 ± 0.082 | 4.523 ± 0.156 | 0.00% ± 0.00 |
| Rank32 × 1.0 | 4 | 0.335 ± 0.034 | 6.185 ± 0.071 | 4.239 ± 0.162 | 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 Bitterness. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Bitterness. 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 Bitterness. 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 Bitterness. 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 0.409 · CE 6.13 · Genuine 4.58 · WER 0.0% | Target 0.351 · CE 6.16 · Genuine 4.33 · WER 0.0% | Target 0.374 · CE 6.21 · Genuine 4.56 · WER 0.0% | Target 0.354 · CE 6.13 · Genuine 4.37 · 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 Bitterness. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Bitterness. 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 Bitterness. 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 Bitterness. 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 0.369 · CE 6.17 · Genuine 4.58 · WER 0.0% | Target 0.328 · CE 6.19 · Genuine 4.37 · WER 0.0% | Target 0.380 · CE 6.06 · Genuine 4.62 · WER 0.0% | Target 0.373 · CE 6.19 · Genuine 4.22 · 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 Bitterness. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Bitterness. 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 Bitterness. 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 Bitterness. 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 0.357 · CE 6.17 · Genuine 4.77 · WER 0.0% | Target 0.356 · CE 6.22 · Genuine 4.56 · WER 0.0% | Target 0.287 · CE 6.19 · Genuine 4.62 · WER 0.0% | Target 0.312 · CE 6.28 · Genuine 4.35 · 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 Bitterness. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Bitterness. 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 Bitterness. 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 Bitterness. 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 0.364 · CE 6.18 · Genuine 4.66 · WER 0.0% | Target 0.298 · CE 6.14 · Genuine 4.18 · WER 0.0% | Target 0.287 · CE 6.25 · Genuine 4.29 · WER 0.0% | Target 0.300 · CE 6.13 · Genuine 4.02 · WER 0.0% |
Fixed speaker reference:
| Condition | n | Target score ↑ | Content Enjoyment ↑ | Genuineness ↑ | WER ↓ |
|---|---|---|---|---|---|
| Merged base, no bucket adapter | 4 | 0.404 ± 0.039 | 6.094 ± 0.060 | 4.299 ± 0.159 | 0.00% ± 0.00 |
| Rank32 × 0.5 | 4 | 0.394 ± 0.031 | 6.126 ± 0.068 | 4.275 ± 0.099 | 0.00% ± 0.00 |
| Rank32 × 0.75 | 4 | 0.378 ± 0.015 | 6.096 ± 0.052 | 4.314 ± 0.159 | 0.00% ± 0.00 |
| Rank32 × 1.0 | 4 | 0.394 ± 0.024 | 6.160 ± 0.019 | 4.303 ± 0.176 | 1.39% ± 2.78 |
| 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 Bitterness. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Bitterness. 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 Bitterness. 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 Bitterness. 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 0.416 · CE 6.11 · Genuine 4.11 · WER 0.0% | Target 0.420 · CE 6.13 · Genuine 4.20 · WER 0.0% | Target 0.372 · CE 6.10 · Genuine 4.35 · WER 0.0% | Target 0.365 · CE 6.17 · Genuine 4.19 · 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 Bitterness. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Bitterness. 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 Bitterness. 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 Bitterness. 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 0.387 · CE 6.02 · Genuine 4.41 · WER 0.0% | Target 0.354 · CE 6.03 · Genuine 4.26 · WER 0.0% | Target 0.401 · CE 6.03 · Genuine 4.39 · WER 0.0% | Target 0.410 · CE 6.17 · Genuine 4.37 · 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 Bitterness. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Bitterness. 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 Bitterness. 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 Bitterness. 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 0.361 · CE 6.08 · Genuine 4.44 · WER 0.0% | Target 0.415 · CE 6.18 · Genuine 4.42 · WER 0.0% | Target 0.368 · CE 6.09 · Genuine 4.08 · WER 0.0% | Target 0.418 · CE 6.13 · Genuine 4.52 · WER 5.6% |
| 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 Bitterness. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Bitterness. 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 Bitterness. 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 Bitterness. 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 0.453 · CE 6.17 · Genuine 4.23 · WER 0.0% | Target 0.387 · CE 6.17 · Genuine 4.22 · WER 0.0% | Target 0.369 · CE 6.16 · Genuine 4.43 · WER 0.0% | Target 0.385 · CE 6.16 · Genuine 4.13 · WER 0.0% |
Fixed speaker reference:
| Condition | n | Target score ↑ | Content Enjoyment ↑ | Genuineness ↑ | WER ↓ |
|---|---|---|---|---|---|
| Merged base, no bucket adapter | 4 | 0.330 ± 0.044 | 6.044 ± 0.044 | 4.259 ± 0.251 | 0.00% ± 0.00 |
| Rank32 × 0.5 | 4 | 0.328 ± 0.047 | 6.140 ± 0.131 | 4.426 ± 0.169 | 0.00% ± 0.00 |
| Rank32 × 0.75 | 4 | 0.323 ± 0.030 | 6.011 ± 0.106 | 4.269 ± 0.337 | 1.67% ± 3.33 |
| Rank32 × 1.0 | 4 | 0.325 ± 0.049 | 6.108 ± 0.051 | 4.302 ± 0.252 | 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 Bitterness. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Bitterness. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La puerta sigue abierta. (Intense Bitterness. 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 Bitterness. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Dime cuándo estás listo. | Target 0.304 · CE 6.01 · Genuine 3.95 · WER 0.0% | Target 0.335 · CE 5.97 · Genuine 4.26 · WER 0.0% | Target 0.309 · CE 5.86 · Genuine 4.03 · WER 6.7% | Target 0.391 · CE 6.08 · Genuine 4.33 · 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 Bitterness. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Bitterness. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La puerta sigue abierta. (Intense Bitterness. 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 Bitterness. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Dime cuándo estás listo. | Target 0.396 · CE 6.05 · Genuine 4.19 · WER 0.0% | Target 0.392 · CE 6.13 · Genuine 4.38 · WER 0.0% | Target 0.367 · CE 6.03 · Genuine 4.01 · WER 0.0% | Target 0.334 · CE 6.07 · Genuine 3.94 · 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 Bitterness. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Bitterness. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] He encontrado la cajita. (Intense Bitterness. 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 Bitterness. 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 0.320 · CE 6.01 · Genuine 4.53 · WER 0.0% | Target 0.292 · CE 6.17 · Genuine 4.41 · WER 0.0% | Target 0.300 · CE 6.08 · Genuine 4.30 · WER 0.0% | Target 0.282 · CE 6.18 · Genuine 4.48 · 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 Bitterness. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Bitterness. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] He encontrado la cajita. (Intense Bitterness. 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 Bitterness. 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 0.302 · CE 6.10 · Genuine 4.37 · WER 0.0% | Target 0.294 · CE 6.29 · Genuine 4.66 · WER 0.0% | Target 0.315 · CE 6.08 · Genuine 4.73 · WER 0.0% | Target 0.294 · CE 6.10 · Genuine 4.46 · WER 0.0% |