← all dimensions · markdown (agent skill) · manual home
Content Appropriateness (3-point Scale) VoiceNet dimension guide
Language & Recordingcode vn_EXPLslot 74
This dimension measures content appropriateness for different audiences (3-point scale). Below: the merge-scaling dose-response for the __high and __low LoRA (laion/moss-voicenet-dimension-loras) at 25/50/75/100/125%, the measured target value at each dose, the top-3 ± correlated EmoNet emotions and VoiceNet dimensions, three audio samples per dose, and a grounded prompting tip. Costs are shifts vs. the 24-sample neutral baseline in genuineness, vocal-burst blend and speech quality.
Content Appropriateness (3-point Scale) controls content appropriateness for different audiences (3-point scale). The __high LoRA shifts it +0.11 at 125% (worst cost: burst-blend -0.33); the __low LoRA shifts it -0.03 at 50% (worst cost: speech quality -0.08). The __high direction is the stronger, more reliable knob here.
▲ HIGH — Content Appropriateness (3-point Scale) __high
Dose-response
| dose | target | Δtarget | Δgenu | Δblend | Δqual |
|---|
| baseline (0%) | 0.026 | — | — | — | — |
| 25% | 0.020 | -0.01 | -0.04 | -0.04 | +0.11 |
| 50% | 0.061 | +0.04 | -0.09 | -0.23 | +0.05 |
| 75% | 0.074 | +0.05 | -0.04 | -0.34 | +0.03 |
| 100% | 0.114 | +0.09 | -0.11 | -0.38 | +0.00 |
| 125% ★ | 0.137 | +0.11 | -0.04 | -0.33 | +0.00 |
EmoNet emotion correlation
Amusement+0.59
Teasing+0.58
Impatience and Irritability+0.53
Longing-0.46
Affection-0.45
Interest-0.37
VoiceNet dimension correlation
Conversational Style+0.69
Arousal Shift+0.62
Tempo+0.55
Monologue Style-0.48
Vocal-burst blend-0.46
Whisper-Talk Style-0.42
Listen — 3 samples per dose
25%measured target: 0.09 · 0.00 · 0.05
50%measured target: 0.00 · 0.06 · 0.03
75%measured target: 0.00 · 0.00 · 0.24
100%measured target: 0.10 · 0.00 · 0.00
125% ★measured target: 0.31 · 0.21 · 0.00
Merging Content Appropriateness (3-point Scale) __high raises the dimension: at the recommended 125% dose the measured Content Appropriateness (3-point Scale) score moves from a 0.03 baseline to 0.14 (a +0.11 shift). The dose-response is clean and monotonic — each step pushes the dimension further. That reach is a steep cost — the biggest hit is to burst-blend (-0.33); across the board genuineness -0.04, burst-blend -0.33, speech quality +0.00. Emotionally it reads as Amusement (r 0.59), Teasing (0.58) and Impatience and Irritability (0.53), and it moves away from Longing, Affection and Interest. On other voice dimensions it co-moves most with Conversational Style (0.69), Arousal Shift (0.62) and Tempo (0.55), while pulling down Monologue Style, Vocal-burst blend and Whisper-Talk Style. Practical call: merge __high at 100–125%; expect Amusement and Conversational Style to rise and burst-blend to fall. If you don’t want Amusement bleeding in, offset it with a prompt cue toward Longing to keep the content appropriateness (3-point scale) shift without the side-effect.
▼ LOW — Content Appropriateness (3-point Scale) __low
Dose-response
| dose | target | Δtarget | Δgenu | Δblend | Δqual |
|---|
| baseline (0%) | 0.026 | — | — | — | — |
| 25% | 0.013 | -0.01 | +0.00 | +0.02 | -0.02 |
| 50% ★ | 0.000 | -0.03 | +0.02 | -0.03 | -0.08 |
| 75% | 0.030 | +0.00 | -0.06 | -0.15 | -0.01 |
| 100% | 0.051 | +0.03 | -0.07 | -0.29 | +0.05 |
| 125% | 0.043 | +0.02 | -0.09 | -0.16 | +0.07 |
EmoNet emotion correlation
Sourness+0.49
Contempt+0.41
Pride+0.39
Longing-0.38
Contemplation-0.31
Infatuation-0.25
VoiceNet dimension correlation
Brightness+0.57
Mask Resonance+0.55
Tempo+0.53
Cognitive Load-0.43
Vocal-burst blend-0.40
Vulnerability-0.37
Listen — 3 samples per dose
25%measured target: 0.00 · 0.00 · 0.09
50% ★measured target: 0.00 · 0.00 · 0.00
75%measured target: 0.00 · 0.24 · 0.00
100%measured target: 0.00 · 0.01 · 0.07
125%measured target: 0.00 · 0.00 · 0.15
Merging Content Appropriateness (3-point Scale) __low pulls down the dimension: at the recommended 50% dose the measured Content Appropriateness (3-point Scale) score moves from a 0.03 baseline to 0.00 (a -0.03 shift). The dose-response is mostly ordered but noisy between doses. That reach is a modest cost — the biggest hit is to speech quality (-0.08); across the board genuineness +0.02, burst-blend -0.03, speech quality -0.08. Emotionally it reads as Sourness (r 0.49), Contempt (0.41) and Pride (0.39), and it moves away from Longing, Contemplation and Infatuation. On other voice dimensions it co-moves most with Brightness (0.57), Mask Resonance (0.55) and Tempo (0.53), while pulling down Cognitive Load, Vocal-burst blend and Vulnerability. Practical call: merge __low at 25–75%; expect Sourness and Brightness to rise and speech quality to fall. If you don’t want Sourness bleeding in, offset it with a prompt cue toward Longing to keep the content appropriateness (3-point scale) shift without the side-effect.