← all dimensions · markdown (agent skill) · manual home

Authoritative Style VoiceNet dimension guide

Speaking Stylecode vn_S_AUTHslot 89
This dimension measures how authoritative and commanding the voice sounds. 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.

Authoritative Style controls how authoritative and commanding the voice sounds. The __high LoRA shifts it +0.63 at 125% (worst cost: burst-blend -0.46); the __low LoRA shifts it -0.01 at 25% (worst cost: burst-blend -0.08). The __high direction is the stronger, more reliable knob here.
▲ HIGH — Authoritative Style __high
Dose-response
dosetargetΔtargetΔgenuΔblendΔqual
baseline (0%)0.032————
25%0.057+0.03-0.04+0.00+0.03
50%0.213+0.18-0.08-0.13+0.07
75%0.482+0.45-0.12-0.37+0.07
100%0.647+0.61-0.13-0.47-0.06
125% ★0.658+0.63-0.13-0.46+0.02
EmoNet emotion correlation
Triumph+0.76
Pride+0.62
Anger+0.56
Contentment-0.74
Relief-0.71
Affection-0.66
VoiceNet dimension correlation
Attack+0.95
Metallic Character+0.94
Arousal+0.93
Vulnerability-0.88
Vocal-burst blend-0.74
Casual Style-0.74
Listen — 3 samples per dose
25%measured target: 0.12 · 0.15 · 0.00
50%measured target: 0.33 · 0.16 · 0.39
75%measured target: 0.27 · 0.60 · 0.33
100%measured target: 0.57 · 0.55 · 0.75
125% ★measured target: 0.58 · 0.56 · 0.80
Merging Authoritative Style __high drives the dimension: at the recommended 125% dose the measured Authoritative Style score moves from a 0.03 baseline to 0.66 (a +0.63 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 (-0.46); across the board genuineness -0.13, burst-blend -0.46, speech quality +0.02. Emotionally it reads as Triumph (r 0.76), Pride (0.62) and Anger (0.56), and it moves away from Contentment, Relief and Affection. On other voice dimensions it co-moves most with Attack (0.95), Metallic Character (0.94) and Arousal (0.93), while pulling down Vulnerability, Vocal-burst blend and Casual Style. Practical call: merge __high at 100–125%; expect Triumph and Attack to rise and burst-blend to fall. If you don’t want Triumph bleeding in, offset it with a prompt cue toward Contentment to keep the authoritative style shift without the side-effect.
▼ LOW — Authoritative Style __low
Dose-response
dosetargetΔtargetΔgenuΔblendΔqual
baseline (0%)0.032————
25% ★0.024-0.01+0.01-0.08+0.02
50%0.026-0.01+0.02+0.12-0.03
75%0.048+0.02+0.05-0.09-0.05
100%0.079+0.05+0.03+0.00-0.07
125%0.183+0.15+0.05-0.08-0.28
EmoNet emotion correlation
Bitterness+0.60
Contempt+0.58
Pride+0.38
Affection-0.48
Relief-0.42
Contentment-0.39
VoiceNet dimension correlation
Attack+0.75
Tempo+0.66
Stance+0.63
Vulnerability-0.48
Vocal-burst blend-0.29
Cognitive Load-0.28
Listen — 3 samples per dose
25% ★measured target: 0.00 · 0.00 · 0.00
50%measured target: 0.02 · 0.00 · 0.00
75%measured target: 0.01 · 0.00 · 0.04
100%measured target: 0.13 · 0.00 · 0.09
125%measured target: 0.44 · 0.13 · 0.14
Merging Authoritative Style __low damps the dimension: at the recommended 25% dose the measured Authoritative Style score moves from a 0.03 baseline to 0.02 (a -0.01 shift). The dose-response is noisy and non-monotonic — the merge is an unreliable knob here. That reach is a modest cost — the biggest hit is to burst-blend (-0.08); across the board genuineness +0.01, burst-blend -0.08, speech quality +0.02. Emotionally it reads as Bitterness (r 0.60), Contempt (0.58) and Pride (0.38), and it moves away from Affection, Relief and Contentment. On other voice dimensions it co-moves most with Attack (0.75), Tempo (0.66) and Stance (0.63), while pulling down Vulnerability, Vocal-burst blend and Cognitive Load. Practical call: merge __low at a light 25–50%; expect Bitterness and Attack to rise and burst-blend to fall. If you don’t want Bitterness bleeding in, offset it with a prompt cue toward Affection to keep the authoritative style shift without the side-effect.
Data: MOSS-VA-v2 base + per-dimension LoRA, neutral text, no reference · 5 merge doses × 8 scored samples/dose · correlations are Pearson r over the sample pool · scaling grid · LoRA repo