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 4.0.
| Held-out condition | n | Target z↑ | Raw head | CE↑ | Genuine↑ | Voice↑ | WER↓ | Valid | QA |
|---|---|---|---|---|---|---|---|---|---|
| standard | 50 | 0.220 | 0.402 | 6.218 | 4.468 | 0.603 | 0.012 | 0.980 | pass |
| manual | 50 | 3.936 | 2.412 | 5.728 | 4.545 | 0.585 | 0.034 | 0.980 | pass |
| adapter | 50 | 4.068 | 2.483 | 5.529 | 4.423 | 0.545 | 0.063 | 0.920 | fail |
| matched_no_adapter | 50 | 4.258 | 2.585 | 5.586 | 4.447 | 0.524 | 0.063 | 0.920 | fail |
Incremental adapter gain versus same-CFG control: -0.189 z;95% cluster interval [-0.510468768891861, 0.12062499579798172]. Versus best manual-only: 0.132 z. Manual-only versus standard: 3.716 z.
| Scale | CFG | Target z | CE | Genuine | WER | Provisional QA |
|---|---|---|---|---|---|---|
| 0.0 | 1.0 | 1.041 | 6.126 | 4.512 | 0.011 | True |
| 0.5 | 1.0 | 1.166 | 6.140 | 4.516 | 0.011 | True |
| 0.75 | 1.0 | 1.260 | 6.153 | 4.575 | 0.006 | True |
| 1.0 | 1.0 | 1.086 | 6.121 | 4.493 | 0.000 | True |
| 0.0 | 1.25 | 1.524 | 6.047 | 4.582 | 0.009 | True |
| 0.5 | 1.25 | 1.630 | 6.087 | 4.584 | 0.003 | True |
| 0.75 | 1.25 | 1.137 | 6.003 | 4.525 | 0.021 | True |
| 1.0 | 1.25 | 1.401 | 6.074 | 4.551 | 0.006 | True |
| 0.0 | 1.5 | 1.774 | 6.026 | 4.581 | 0.006 | True |
| 0.5 | 1.5 | 1.634 | 6.058 | 4.546 | 0.006 | True |
| 0.75 | 1.5 | 2.076 | 6.029 | 4.588 | 0.006 | True |
| 1.0 | 1.5 | 1.785 | 6.031 | 4.579 | 0.005 | True |
| 0.0 | 2.0 | 2.533 | 5.951 | 4.612 | 0.033 | True |
| 0.5 | 2.0 | 2.754 | 5.930 | 4.613 | 0.013 | True |
| 0.75 | 2.0 | 2.642 | 5.926 | 4.607 | 0.009 | True |
| 1.0 | 2.0 | 2.841 | 5.939 | 4.589 | 0.006 | True |
| 0.0 | 2.5 | 3.505 | 5.904 | 4.644 | 0.022 | True |
| 0.5 | 2.5 | 3.268 | 5.758 | 4.585 | 0.014 | True |
| 0.75 | 2.5 | 3.454 | 5.822 | 4.610 | 0.011 | True |
| 1.0 | 2.5 | 3.385 | 5.686 | 4.596 | 0.011 | True |
| 0.0 | 3.0 | 3.972 | 5.698 | 4.620 | 0.009 | True |
| 0.5 | 3.0 | 3.640 | 5.642 | 4.582 | 0.006 | True |
| 0.75 | 3.0 | 3.871 | 5.706 | 4.586 | 0.011 | True |
| 1.0 | 3.0 | 3.356 | 5.643 | 4.526 | 0.006 | True |
| 0.0 | 4.0 | 4.495 | 5.546 | 4.537 | 0.003 | True |
| 0.5 | 4.0 | 4.389 | 5.547 | 4.543 | 0.008 | True |
| 0.75 | 4.0 | 4.514 | 5.471 | 4.518 | 0.025 | True |
| 1.0 | 4.0 | 4.206 | 5.410 | 4.462 | 0.023 | True |
| 0.0 | 5.0 | 4.853 | 5.368 | 4.452 | 0.073 | False |
| 0.5 | 5.0 | 4.836 | 5.242 | 4.421 | 0.049 | True |
| 0.75 | 5.0 | 4.770 | 5.294 | 4.379 | 0.028 | True |
| 1.0 | 5.0 | 4.719 | 5.322 | 4.401 | 0.056 | False |
direct_manual_recipe
{
"id": "manual-1711f3012766",
"general": "Embody exhaustion as a systemic drain, where the voice sounds hollowed out and brittle. Prioritize texture over volume.",
"inline": [
"Begin with a breathy onset, implying the lungs are incapable of a full, robust inhalation before speaking.",
"Introduce subtle vocal fry or rasping texture on key consonants to signify deep physical strain.",
"Deliver the last phrase with a fragile, almost cracking quality, as if the vocal apparatus is on the verge of breaking."
],
"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 Fatigue Exhaustion. 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: {"en": 42, "nl": 44, "de": 34, "es": 36, "fr": 41}. 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 | 0.687 ± 0.225 | 6.147 ± 0.151 | 4.440 ± 0.186 | 0.31% ± 1.40 |
| Rank32 × 0.5 | 20 | 0.702 ± 0.237 | 6.140 ± 0.117 | 4.462 ± 0.217 | 0.00% ± 0.00 |
| Rank32 × 0.75 | 20 | 0.686 ± 0.192 | 6.173 ± 0.112 | 4.463 ± 0.224 | 0.28% ± 1.24 |
| Rank32 × 1.0 | 20 | 0.710 ± 0.245 | 6.131 ± 0.144 | 4.397 ± 0.199 | 0.59% ± 1.82 |
Fixed speaker reference:
| Condition | n | Target score ↑ | Content Enjoyment ↑ | Genuineness ↑ | WER ↓ |
|---|---|---|---|---|---|
| Merged base, no bucket adapter | 4 | 0.575 ± 0.161 | 6.279 ± 0.110 | 4.335 ± 0.224 | 1.56% ± 3.12 |
| Rank32 × 0.5 | 4 | 0.544 ± 0.083 | 6.265 ± 0.084 | 4.385 ± 0.226 | 0.00% ± 0.00 |
| Rank32 × 0.75 | 4 | 0.495 ± 0.122 | 6.227 ± 0.094 | 4.252 ± 0.251 | 0.00% ± 0.00 |
| Rank32 × 1.0 | 4 | 0.506 ± 0.164 | 6.213 ± 0.079 | 4.335 ± 0.200 | 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 Fatigue Exhaustion. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Fatigue Exhaustion. 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 Fatigue Exhaustion. 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 Fatigue Exhaustion. 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 0.747 · CE 6.33 · Genuine 4.36 · WER 0.0% | Target 0.565 · CE 6.20 · Genuine 4.29 · WER 0.0% | Target 0.452 · CE 6.26 · Genuine 3.99 · WER 0.0% | Target 0.425 · CE 6.13 · Genuine 4.24 · 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 Fatigue Exhaustion. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Fatigue Exhaustion. 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 Fatigue Exhaustion. 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 Fatigue Exhaustion. 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 0.679 · CE 6.12 · Genuine 4.02 · WER 0.0% | Target 0.645 · CE 6.24 · Genuine 4.13 · WER 0.0% | Target 0.675 · CE 6.10 · Genuine 4.14 · WER 0.0% | Target 0.750 · CE 6.17 · Genuine 4.21 · 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 Fatigue Exhaustion. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Fatigue Exhaustion. 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 Fatigue Exhaustion. 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 Fatigue Exhaustion. 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 0.451 · CE 6.30 · Genuine 4.56 · WER 0.0% | Target 0.446 · CE 6.24 · Genuine 4.46 · WER 0.0% | Target 0.448 · CE 6.32 · Genuine 4.58 · WER 0.0% | Target 0.446 · CE 6.31 · Genuine 4.63 · 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 Fatigue Exhaustion. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Fatigue Exhaustion. 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 Fatigue Exhaustion. 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 Fatigue Exhaustion. 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 0.426 · CE 6.36 · Genuine 4.40 · WER 6.2% | Target 0.519 · CE 6.39 · Genuine 4.66 · WER 0.0% | Target 0.404 · CE 6.22 · Genuine 4.29 · WER 0.0% | Target 0.401 · CE 6.25 · Genuine 4.26 · WER 0.0% |
Fixed speaker reference:
| Condition | n | Target score ↑ | Content Enjoyment ↑ | Genuineness ↑ | WER ↓ |
|---|---|---|---|---|---|
| Merged base, no bucket adapter | 4 | 0.582 ± 0.304 | 6.257 ± 0.071 | 4.391 ± 0.203 | 0.00% ± 0.00 |
| Rank32 × 0.5 | 4 | 0.660 ± 0.341 | 6.203 ± 0.146 | 4.467 ± 0.252 | 0.00% ± 0.00 |
| Rank32 × 0.75 | 4 | 0.719 ± 0.218 | 6.289 ± 0.063 | 4.623 ± 0.129 | 0.00% ± 0.00 |
| Rank32 × 1.0 | 4 | 0.641 ± 0.339 | 6.289 ± 0.035 | 4.390 ± 0.225 | 1.56% ± 3.12 |
| 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 Fatigue Exhaustion. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Fatigue Exhaustion. 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 Fatigue Exhaustion. 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 Fatigue Exhaustion. 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 0.639 · CE 6.16 · Genuine 4.43 · WER 0.0% | Target 0.987 · CE 6.13 · Genuine 4.63 · WER 0.0% | Target 0.849 · CE 6.23 · Genuine 4.48 · WER 0.0% | Target 0.871 · CE 6.28 · Genuine 4.48 · 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 Fatigue Exhaustion. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Fatigue Exhaustion. 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 Fatigue Exhaustion. 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 Fatigue Exhaustion. 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 0.988 · CE 6.26 · Genuine 4.66 · WER 0.0% | Target 0.921 · CE 6.07 · Genuine 4.48 · WER 0.0% | Target 0.957 · CE 6.25 · Genuine 4.78 · WER 0.0% | Target 0.992 · CE 6.34 · Genuine 4.66 · 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 Fatigue Exhaustion. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Fatigue Exhaustion. 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 Fatigue Exhaustion. 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 Fatigue Exhaustion. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Wir können gemeinsam hineinschauen. | Target 0.383 · CE 6.28 · Genuine 4.19 · WER 0.0% | Target 0.392 · CE 6.40 · Genuine 4.65 · WER 0.0% | Target 0.571 · CE 6.37 · Genuine 4.57 · WER 0.0% | Target 0.358 · CE 6.27 · Genuine 4.26 · WER 6.2% |
| 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 Fatigue Exhaustion. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Fatigue Exhaustion. 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 Fatigue Exhaustion. 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 Fatigue Exhaustion. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Wir können gemeinsam hineinschauen. | Target 0.318 · CE 6.32 · Genuine 4.28 · WER 0.0% | Target 0.342 · CE 6.21 · Genuine 4.11 · WER 0.0% | Target 0.501 · CE 6.30 · Genuine 4.66 · WER 0.0% | Target 0.342 · CE 6.27 · Genuine 4.16 · WER 0.0% |
Fixed speaker reference:
| Condition | n | Target score ↑ | Content Enjoyment ↑ | Genuineness ↑ | WER ↓ |
|---|---|---|---|---|---|
| Merged base, no bucket adapter | 4 | 0.583 ± 0.198 | 6.100 ± 0.131 | 4.478 ± 0.170 | 0.00% ± 0.00 |
| Rank32 × 0.5 | 4 | 0.824 ± 0.165 | 6.042 ± 0.066 | 4.620 ± 0.118 | 0.00% ± 0.00 |
| Rank32 × 0.75 | 4 | 0.697 ± 0.099 | 6.104 ± 0.090 | 4.573 ± 0.100 | 0.00% ± 0.00 |
| Rank32 × 1.0 | 4 | 0.729 ± 0.190 | 6.070 ± 0.101 | 4.496 ± 0.170 | 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 Fatigue Exhaustion. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Fatigue Exhaustion. 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 Fatigue Exhaustion. 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 Fatigue Exhaustion. 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 0.639 · CE 6.12 · Genuine 4.72 · WER 0.0% | Target 0.950 · CE 6.09 · Genuine 4.70 · WER 0.0% | Target 0.739 · CE 6.17 · Genuine 4.47 · WER 0.0% | Target 0.940 · CE 6.19 · Genuine 4.60 · 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 Fatigue Exhaustion. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Fatigue Exhaustion. 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 Fatigue Exhaustion. 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 Fatigue Exhaustion. 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 0.840 · CE 5.94 · Genuine 4.36 · WER 0.0% | Target 0.763 · CE 6.00 · Genuine 4.44 · WER 0.0% | Target 0.809 · CE 6.11 · Genuine 4.51 · WER 0.0% | Target 0.801 · CE 6.09 · Genuine 4.52 · 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 Fatigue Exhaustion. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Fatigue Exhaustion. 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 Fatigue Exhaustion. 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 Fatigue Exhaustion. 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 0.427 · CE 6.08 · Genuine 4.47 · WER 0.0% | Target 0.965 · CE 5.97 · Genuine 4.66 · WER 0.0% | Target 0.582 · CE 6.17 · Genuine 4.67 · WER 0.0% | Target 0.685 · CE 5.94 · Genuine 4.25 · 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 Fatigue Exhaustion. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Fatigue Exhaustion. 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 Fatigue Exhaustion. 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 Fatigue Exhaustion. 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 0.427 · CE 6.26 · Genuine 4.36 · WER 0.0% | Target 0.619 · CE 6.11 · Genuine 4.68 · WER 0.0% | Target 0.658 · CE 5.97 · Genuine 4.65 · WER 0.0% | Target 0.490 · CE 6.06 · Genuine 4.62 · WER 0.0% |
Fixed speaker reference:
| Condition | n | Target score ↑ | Content Enjoyment ↑ | Genuineness ↑ | WER ↓ |
|---|---|---|---|---|---|
| Merged base, no bucket adapter | 4 | 0.912 ± 0.104 | 6.116 ± 0.115 | 4.430 ± 0.095 | 0.00% ± 0.00 |
| Rank32 × 0.5 | 4 | 0.791 ± 0.219 | 6.090 ± 0.084 | 4.330 ± 0.121 | 0.00% ± 0.00 |
| Rank32 × 0.75 | 4 | 0.828 ± 0.196 | 6.174 ± 0.117 | 4.350 ± 0.202 | 1.39% ± 2.78 |
| Rank32 × 1.0 | 4 | 0.893 ± 0.087 | 6.038 ± 0.177 | 4.286 ± 0.151 | 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 Fatigue Exhaustion. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Fatigue Exhaustion. 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 Fatigue Exhaustion. 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 Fatigue Exhaustion. 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 0.974 · CE 5.94 · Genuine 4.32 · WER 0.0% | Target 0.986 · CE 6.12 · Genuine 4.48 · WER 0.0% | Target 0.789 · CE 6.03 · Genuine 4.08 · WER 0.0% | Target 0.785 · CE 5.90 · Genuine 4.07 · 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 Fatigue Exhaustion. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Fatigue Exhaustion. 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 Fatigue Exhaustion. 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 Fatigue Exhaustion. 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 0.966 · CE 6.17 · Genuine 4.39 · WER 0.0% | Target 0.936 · CE 5.97 · Genuine 4.19 · WER 0.0% | Target 0.566 · CE 6.28 · Genuine 4.38 · WER 0.0% | Target 0.996 · CE 5.90 · Genuine 4.31 · 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 Fatigue Exhaustion. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Fatigue Exhaustion. 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 Fatigue Exhaustion. 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 Fatigue Exhaustion. 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 0.757 · CE 6.18 · Genuine 4.53 · WER 0.0% | Target 0.737 · CE 6.12 · Genuine 4.33 · WER 0.0% | Target 0.961 · CE 6.13 · Genuine 4.38 · WER 0.0% | Target 0.904 · CE 6.09 · Genuine 4.37 · 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 Fatigue Exhaustion. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Fatigue Exhaustion. 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 Fatigue Exhaustion. 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 Fatigue Exhaustion. 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 0.951 · CE 6.17 · Genuine 4.48 · WER 0.0% | Target 0.505 · CE 6.16 · Genuine 4.32 · WER 0.0% | Target 0.995 · CE 6.26 · Genuine 4.57 · WER 5.6% | Target 0.886 · CE 6.27 · Genuine 4.40 · WER 5.6% |
Fixed speaker reference:
| Condition | n | Target score ↑ | Content Enjoyment ↑ | Genuineness ↑ | WER ↓ |
|---|---|---|---|---|---|
| Merged base, no bucket adapter | 4 | 0.782 ± 0.161 | 5.982 ± 0.127 | 4.567 ± 0.221 | 0.00% ± 0.00 |
| Rank32 × 0.5 | 4 | 0.690 ± 0.300 | 6.100 ± 0.050 | 4.510 ± 0.295 | 0.00% ± 0.00 |
| Rank32 × 0.75 | 4 | 0.693 ± 0.212 | 6.072 ± 0.054 | 4.515 ± 0.237 | 0.00% ± 0.00 |
| Rank32 × 1.0 | 4 | 0.783 ± 0.282 | 6.042 ± 0.118 | 4.478 ± 0.257 | 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 Fatigue Exhaustion. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Fatigue Exhaustion. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La puerta sigue abierta. (Intense Fatigue Exhaustion. 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 Fatigue Exhaustion. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Dime cuándo estás listo. | Target 0.998 · CE 5.80 · Genuine 4.53 · WER 0.0% | Target 0.921 · CE 6.09 · Genuine 4.52 · WER 0.0% | Target 0.990 · CE 6.04 · Genuine 4.42 · WER 0.0% | Target 0.972 · CE 6.14 · Genuine 4.42 · 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 Fatigue Exhaustion. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Fatigue Exhaustion. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La puerta sigue abierta. (Intense Fatigue Exhaustion. 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 Fatigue Exhaustion. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Dime cuándo estás listo. | Target 0.612 · CE 5.98 · Genuine 4.27 · WER 0.0% | Target 0.445 · CE 6.07 · Genuine 4.09 · WER 0.0% | Target 0.682 · CE 6.02 · Genuine 4.22 · WER 0.0% | Target 0.958 · CE 5.87 · Genuine 4.17 · 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 Fatigue Exhaustion. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Fatigue Exhaustion. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] He encontrado la cajita. (Intense Fatigue Exhaustion. 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 Fatigue Exhaustion. 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 0.734 · CE 6.05 · Genuine 4.71 · WER 0.0% | Target 0.417 · CE 6.07 · Genuine 4.67 · WER 0.0% | Target 0.599 · CE 6.14 · Genuine 4.72 · WER 0.0% | Target 0.371 · CE 6.08 · Genuine 4.52 · 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 Fatigue Exhaustion. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Fatigue Exhaustion. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] He encontrado la cajita. (Intense Fatigue Exhaustion. 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 Fatigue Exhaustion. 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 0.785 · CE 6.09 · Genuine 4.76 · WER 0.0% | Target 0.977 · CE 6.17 · Genuine 4.76 · WER 0.0% | Target 0.499 · CE 6.09 · Genuine 4.70 · WER 0.0% | Target 0.832 · CE 6.08 · Genuine 4.80 · WER 0.0% |