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 | -0.546 | 0.019 | 6.277 | 4.345 | 0.572 | 0.007 | 1.000 | pass |
| manual | 50 | -0.452 | 0.061 | 6.051 | 4.673 | 0.618 | 0.020 | 0.980 | pass |
| adapter | 50 | -0.494 | 0.042 | 6.097 | 4.602 | 0.634 | 0.020 | 1.000 | pass |
| matched_no_adapter | 50 | -0.466 | 0.054 | 6.085 | 4.617 | 0.637 | 0.033 | 0.960 | pass |
Incremental adapter gain versus same-CFG control: -0.028 z;95% cluster interval [-0.07065430162096883, 0.003574216015757249]. Versus best manual-only: -0.042 z. Manual-only versus standard: 0.094 z.
| Scale | CFG | Target z | CE | Genuine | WER | Provisional QA |
|---|---|---|---|---|---|---|
| 0.0 | 1.0 | -0.507 | 6.127 | 4.520 | 0.006 | True |
| 0.5 | 1.0 | -0.519 | 6.150 | 4.530 | 0.000 | True |
| 0.75 | 1.0 | -0.492 | 6.136 | 4.499 | 0.003 | True |
| 1.0 | 1.0 | -0.512 | 6.127 | 4.536 | 0.000 | True |
| 0.0 | 1.25 | -0.524 | 6.124 | 4.545 | 0.008 | True |
| 0.5 | 1.25 | -0.523 | 6.114 | 4.478 | 0.015 | True |
| 0.75 | 1.25 | -0.523 | 6.122 | 4.481 | 0.014 | True |
| 1.0 | 1.25 | -0.519 | 6.156 | 4.499 | 0.000 | True |
| 0.0 | 1.5 | -0.524 | 6.096 | 4.545 | 0.000 | True |
| 0.5 | 1.5 | -0.528 | 6.125 | 4.554 | 0.003 | True |
| 0.75 | 1.5 | -0.510 | 6.097 | 4.544 | 0.003 | True |
| 1.0 | 1.5 | -0.515 | 6.145 | 4.572 | 0.005 | True |
| 0.0 | 2.0 | -0.520 | 6.055 | 4.621 | 0.011 | True |
| 0.5 | 2.0 | -0.508 | 6.079 | 4.610 | 0.011 | True |
| 0.75 | 2.0 | -0.513 | 6.014 | 4.621 | 0.014 | True |
| 1.0 | 2.0 | -0.537 | 6.085 | 4.554 | 0.011 | True |
| 0.0 | 2.5 | -0.521 | 6.054 | 4.637 | 0.006 | True |
| 0.5 | 2.5 | -0.520 | 6.042 | 4.599 | 0.011 | True |
| 0.75 | 2.5 | -0.514 | 6.038 | 4.607 | 0.012 | True |
| 1.0 | 2.5 | -0.517 | 6.022 | 4.635 | 0.011 | True |
| 0.0 | 3.0 | -0.518 | 5.960 | 4.635 | 0.014 | True |
| 0.5 | 3.0 | -0.514 | 6.028 | 4.631 | 0.011 | True |
| 0.75 | 3.0 | -0.496 | 5.969 | 4.631 | 0.012 | True |
| 1.0 | 3.0 | -0.461 | 6.014 | 4.658 | 0.017 | True |
| 0.0 | 4.0 | -0.490 | 5.932 | 4.692 | 0.015 | True |
| 0.5 | 4.0 | -0.476 | 5.971 | 4.673 | 0.028 | True |
| 0.75 | 4.0 | -0.493 | 5.975 | 4.678 | 0.062 | True |
| 1.0 | 4.0 | -0.472 | 5.918 | 4.648 | 0.022 | True |
| 0.0 | 5.0 | -0.502 | 5.868 | 4.687 | 0.044 | True |
| 0.5 | 5.0 | -0.513 | 5.922 | 4.665 | 0.022 | True |
| 0.75 | 5.0 | -0.503 | 5.896 | 4.697 | 0.014 | True |
| 1.0 | 5.0 | -0.462 | 5.907 | 4.700 | 0.040 | True |
direct_manual_recipe
{
"id": "manual-21ce2b7b89e4",
"general": "Self-conscious discomfort; a small awkward smile, hesitant timing and a voice wanting to hide. Establish it from the first phrase. Self-conscious discomfort; a small awkward smile, hesitant timing and a voice wanting to hide. Sustain it through the thought. Self-conscious discomfort; a small awkward smile, hesitant timing and a voice wanting to hide. Keep it audible in the final sentence. Begin with the feeling held back, then let it leak through and become unmistakable.",
"inline": [
"Self-conscious discomfort; a small awkward smile, hesitant timing and a voice wanting to hide. Establish it from the first phrase. Begin with the feeling held back, then let it leak through and become unmistakable.",
"Self-conscious discomfort; a small awkward smile, hesitant timing and a voice wanting to hide. Sustain it through the thought. Begin with the feeling held back, then let it leak through and become unmistakable.",
"Self-conscious discomfort; a small awkward smile, hesitant timing and a voice wanting to hide. Keep it audible in the final sentence. Begin with the feeling held back, then let it leak through and become unmistakable."
],
"cfg": 3,
"recipe": "P06",
"strategy": "initial-physical-directing",
"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 Embarrassment. 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": 11, "en": 12, "es": 4, "fr": 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 | 0.304 ± 0.015 | 6.201 ± 0.132 | 4.401 ± 0.214 | 0.28% ± 1.24 |
| Rank32 × 0.5 | 20 | 0.302 ± 0.018 | 6.189 ± 0.137 | 4.334 ± 0.257 | 0.53% ± 2.35 |
| Rank32 × 0.75 | 20 | 0.300 ± 0.016 | 6.185 ± 0.129 | 4.331 ± 0.204 | 0.00% ± 0.00 |
| Rank32 × 1.0 | 20 | 0.303 ± 0.012 | 6.192 ± 0.150 | 4.393 ± 0.205 | 0.00% ± 0.00 |
Fixed speaker reference:
| Condition | n | Target score ↑ | Content Enjoyment ↑ | Genuineness ↑ | WER ↓ |
|---|---|---|---|---|---|
| Merged base, no bucket adapter | 4 | 0.304 ± 0.016 | 6.303 ± 0.142 | 4.285 ± 0.262 | 0.00% ± 0.00 |
| Rank32 × 0.5 | 4 | 0.306 ± 0.015 | 6.277 ± 0.129 | 4.106 ± 0.298 | 0.00% ± 0.00 |
| Rank32 × 0.75 | 4 | 0.305 ± 0.013 | 6.239 ± 0.151 | 4.156 ± 0.313 | 0.00% ± 0.00 |
| Rank32 × 1.0 | 4 | 0.302 ± 0.014 | 6.342 ± 0.096 | 4.206 ± 0.266 | 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 Embarrassment. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Embarrassment. 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 Embarrassment. 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 Embarrassment. 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 0.315 · CE 6.30 · Genuine 4.38 · WER 0.0% | Target 0.325 · CE 6.25 · Genuine 4.12 · WER 0.0% | Target 0.314 · CE 6.04 · Genuine 3.72 · WER 0.0% | Target 0.315 · CE 6.28 · Genuine 3.88 · 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 Embarrassment. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Embarrassment. 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 Embarrassment. 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 Embarrassment. 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 0.320 · CE 6.17 · Genuine 3.92 · WER 0.0% | Target 0.310 · CE 6.16 · Genuine 3.70 · WER 0.0% | Target 0.316 · CE 6.28 · Genuine 4.13 · WER 0.0% | Target 0.309 · CE 6.25 · Genuine 4.11 · WER 0.0% |
| Seed 1701 I found the little box. It was beside the old window. We can look inside together. Exact GENERAL / SCRIPT promptGENERAL: One adult speaker in a clean recording. Intense Embarrassment. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Embarrassment. 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 Embarrassment. 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 Embarrassment. 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 0.295 · CE 6.25 · Genuine 4.31 · WER 0.0% | Target 0.294 · CE 6.46 · Genuine 4.42 · WER 0.0% | Target 0.287 · CE 6.40 · Genuine 4.36 · WER 0.0% | Target 0.300 · CE 6.38 · Genuine 4.37 · 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 Embarrassment. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Embarrassment. 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 Embarrassment. 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 Embarrassment. 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 0.286 · CE 6.50 · Genuine 4.53 · WER 0.0% | Target 0.294 · CE 6.25 · Genuine 4.18 · WER 0.0% | Target 0.303 · CE 6.24 · Genuine 4.41 · WER 0.0% | Target 0.283 · CE 6.46 · Genuine 4.47 · WER 0.0% |
Fixed speaker reference:
| Condition | n | Target score ↑ | Content Enjoyment ↑ | Genuineness ↑ | WER ↓ |
|---|---|---|---|---|---|
| Merged base, no bucket adapter | 4 | 0.304 ± 0.025 | 6.305 ± 0.150 | 4.457 ± 0.220 | 0.00% ± 0.00 |
| Rank32 × 0.5 | 4 | 0.311 ± 0.022 | 6.257 ± 0.179 | 4.432 ± 0.121 | 0.00% ± 0.00 |
| Rank32 × 0.75 | 4 | 0.305 ± 0.022 | 6.305 ± 0.181 | 4.467 ± 0.179 | 0.00% ± 0.00 |
| Rank32 × 1.0 | 4 | 0.307 ± 0.019 | 6.303 ± 0.086 | 4.404 ± 0.143 | 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 Embarrassment. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Embarrassment. 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 Embarrassment. 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 Embarrassment. 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 0.318 · CE 6.19 · Genuine 4.66 · WER 0.0% | Target 0.329 · CE 6.10 · Genuine 4.48 · WER 0.0% | Target 0.321 · CE 6.23 · Genuine 4.49 · WER 0.0% | Target 0.327 · CE 6.18 · Genuine 4.29 · 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 Embarrassment. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Embarrassment. 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 Embarrassment. 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 Embarrassment. 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 0.332 · CE 6.23 · Genuine 4.51 · WER 0.0% | Target 0.327 · CE 6.10 · Genuine 4.53 · WER 0.0% | Target 0.327 · CE 6.09 · Genuine 4.65 · WER 0.0% | Target 0.319 · CE 6.31 · Genuine 4.55 · 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 Embarrassment. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Embarrassment. 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 Embarrassment. 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 Embarrassment. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Wir können gemeinsam hineinschauen. | Target 0.282 · CE 6.28 · Genuine 4.14 · WER 0.0% | Target 0.283 · CE 6.40 · Genuine 4.26 · WER 0.0% | Target 0.283 · CE 6.51 · Genuine 4.22 · WER 0.0% | Target 0.292 · CE 6.38 · Genuine 4.51 · 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 Embarrassment. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Embarrassment. 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 Embarrassment. 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 Embarrassment. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Wir können gemeinsam hineinschauen. | Target 0.286 · CE 6.52 · Genuine 4.52 · WER 0.0% | Target 0.303 · CE 6.42 · Genuine 4.46 · WER 0.0% | Target 0.290 · CE 6.39 · Genuine 4.51 · WER 0.0% | Target 0.288 · CE 6.34 · Genuine 4.27 · WER 0.0% |
Fixed speaker reference:
| Condition | n | Target score ↑ | Content Enjoyment ↑ | Genuineness ↑ | WER ↓ |
|---|---|---|---|---|---|
| Merged base, no bucket adapter | 4 | 0.299 ± 0.010 | 6.128 ± 0.077 | 4.573 ± 0.115 | 1.39% ± 2.78 |
| Rank32 × 0.5 | 4 | 0.291 ± 0.011 | 6.176 ± 0.135 | 4.445 ± 0.163 | 0.00% ± 0.00 |
| Rank32 × 0.75 | 4 | 0.295 ± 0.010 | 6.116 ± 0.058 | 4.407 ± 0.089 | 0.00% ± 0.00 |
| Rank32 × 1.0 | 4 | 0.299 ± 0.003 | 6.146 ± 0.096 | 4.515 ± 0.095 | 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 Embarrassment. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Embarrassment. 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 Embarrassment. 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 Embarrassment. 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 0.287 · CE 6.09 · Genuine 4.47 · WER 5.6% | Target 0.296 · CE 6.33 · Genuine 4.62 · WER 0.0% | Target 0.302 · CE 6.20 · Genuine 4.41 · WER 0.0% | Target 0.299 · CE 6.09 · Genuine 4.41 · 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 Embarrassment. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Embarrassment. 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 Embarrassment. 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 Embarrassment. 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 0.299 · CE 6.13 · Genuine 4.49 · WER 0.0% | Target 0.300 · CE 6.01 · Genuine 4.29 · WER 0.0% | Target 0.306 · CE 6.08 · Genuine 4.47 · WER 0.0% | Target 0.300 · CE 6.28 · Genuine 4.46 · 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 Embarrassment. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Embarrassment. 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 Embarrassment. 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 Embarrassment. 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 0.311 · CE 6.24 · Genuine 4.72 · WER 0.0% | Target 0.290 · CE 6.13 · Genuine 4.55 · WER 0.0% | Target 0.287 · CE 6.12 · Genuine 4.28 · WER 0.0% | Target 0.295 · CE 6.07 · Genuine 4.58 · WER 0.0% |
| Seed 1702 J’ai trouvé la petite boîte. Elle était près de la vieille fenêtre. Nous pouvons regarder ensemble ce qu’elle contient. Exact GENERAL / SCRIPT promptGENERAL: One adult speaker in a clean recording. Intense Embarrassment. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Embarrassment. 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 Embarrassment. 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 Embarrassment. 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 0.298 · CE 6.06 · Genuine 4.61 · WER 0.0% | Target 0.276 · CE 6.22 · Genuine 4.32 · WER 0.0% | Target 0.287 · CE 6.07 · Genuine 4.47 · WER 0.0% | Target 0.302 · CE 6.13 · Genuine 4.61 · WER 0.0% |
No accepted bucket-training example in this language: transfer evaluation only.
Fixed speaker reference:
| Condition | n | Target score ↑ | Content Enjoyment ↑ | Genuineness ↑ | WER ↓ |
|---|---|---|---|---|---|
| Merged base, no bucket adapter | 4 | 0.314 ± 0.012 | 6.166 ± 0.053 | 4.369 ± 0.170 | 0.00% ± 0.00 |
| Rank32 × 0.5 | 4 | 0.311 ± 0.021 | 6.138 ± 0.087 | 4.374 ± 0.138 | 2.63% ± 5.26 |
| Rank32 × 0.75 | 4 | 0.306 ± 0.012 | 6.112 ± 0.065 | 4.231 ± 0.051 | 0.00% ± 0.00 |
| Rank32 × 1.0 | 4 | 0.310 ± 0.010 | 6.132 ± 0.059 | 4.369 ± 0.178 | 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 Embarrassment. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Embarrassment. 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 Embarrassment. 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 Embarrassment. 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 0.326 · CE 6.24 · Genuine 4.41 · WER 0.0% | Target 0.342 · CE 6.04 · Genuine 4.23 · WER 10.5% | Target 0.315 · CE 6.10 · Genuine 4.19 · WER 0.0% | Target 0.323 · CE 6.14 · Genuine 4.14 · 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 Embarrassment. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Embarrassment. 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 Embarrassment. 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 Embarrassment. 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 0.322 · CE 6.12 · Genuine 4.53 · WER 0.0% | Target 0.303 · CE 6.11 · Genuine 4.39 · WER 0.0% | Target 0.302 · CE 6.09 · Genuine 4.31 · WER 0.0% | Target 0.309 · CE 6.05 · Genuine 4.32 · 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 Embarrassment. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Embarrassment. 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 Embarrassment. 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 Embarrassment. 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 0.306 · CE 6.17 · Genuine 4.41 · WER 0.0% | Target 0.303 · CE 6.16 · Genuine 4.32 · WER 0.0% | Target 0.315 · CE 6.05 · Genuine 4.21 · WER 0.0% | Target 0.299 · CE 6.17 · Genuine 4.47 · 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 Embarrassment. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Embarrassment. 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 Embarrassment. 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 Embarrassment. 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 0.303 · CE 6.13 · Genuine 4.13 · WER 0.0% | Target 0.295 · CE 6.25 · Genuine 4.56 · WER 0.0% | Target 0.290 · CE 6.21 · Genuine 4.22 · WER 0.0% | Target 0.311 · CE 6.17 · Genuine 4.55 · WER 0.0% |
Fixed speaker reference:
| Condition | n | Target score ↑ | Content Enjoyment ↑ | Genuineness ↑ | WER ↓ |
|---|---|---|---|---|---|
| Merged base, no bucket adapter | 4 | 0.297 ± 0.010 | 6.102 ± 0.097 | 4.322 ± 0.245 | 0.00% ± 0.00 |
| Rank32 × 0.5 | 4 | 0.294 ± 0.016 | 6.096 ± 0.116 | 4.313 ± 0.409 | 0.00% ± 0.00 |
| Rank32 × 0.75 | 4 | 0.290 ± 0.020 | 6.156 ± 0.072 | 4.394 ± 0.176 | 0.00% ± 0.00 |
| Rank32 × 1.0 | 4 | 0.299 ± 0.010 | 6.037 ± 0.164 | 4.473 ± 0.242 | 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 Embarrassment. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Embarrassment. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La puerta sigue abierta. (Intense Embarrassment. 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 Embarrassment. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Dime cuándo estás listo. | Target 0.300 · CE 6.17 · Genuine 4.31 · WER 0.0% | Target 0.300 · CE 5.94 · Genuine 3.94 · WER 0.0% | Target 0.290 · CE 6.11 · Genuine 4.47 · WER 0.0% | Target 0.295 · CE 5.95 · Genuine 4.37 · 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 Embarrassment. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Embarrassment. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] La puerta sigue abierta. (Intense Embarrassment. 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 Embarrassment. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] Dime cuándo estás listo. | Target 0.309 · CE 6.01 · Genuine 4.13 · WER 0.0% | Target 0.314 · CE 6.09 · Genuine 4.23 · WER 0.0% | Target 0.319 · CE 6.09 · Genuine 4.13 · WER 0.0% | Target 0.313 · CE 5.85 · Genuine 4.19 · 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 Embarrassment. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Embarrassment. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] He encontrado la cajita. (Intense Embarrassment. 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 Embarrassment. 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 0.286 · CE 6.02 · Genuine 4.17 · WER 0.0% | Target 0.284 · CE 6.15 · Genuine 4.18 · WER 0.0% | Target 0.274 · CE 6.25 · Genuine 4.49 · WER 0.0% | Target 0.296 · CE 6.17 · Genuine 4.58 · 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 Embarrassment. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word. SCRIPT: (Intense Embarrassment. Make the requested delivery clearly audible, not merely implied by the words. Preserve every word.) [3.2 seconds duration] He encontrado la cajita. (Intense Embarrassment. 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 Embarrassment. 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 0.294 · CE 6.20 · Genuine 4.67 · WER 0.0% | Target 0.279 · CE 6.21 · Genuine 4.90 · WER 0.0% | Target 0.279 · CE 6.17 · Genuine 4.48 · WER 0.0% | Target 0.291 · CE 6.18 · Genuine 4.75 · WER 0.0% |