10-generation evolution of prompt + LoRA-strength + sampling to maximize Concentration. Fitness = mean over 8 candidates. Best evolved 0.217 vs prompt-without-LoRA 0.019 (LoRA lift +0.198) vs neutral 0.108; winning λ≈1.75.
Method & reproducibility — how the evolution ran
Base model:laion/moss-tts-local-transformer-4.55b-voice-acting-v2 · LoRA: the matching v3 emotion LoRA (TTS-AGI/moss-emotion-loras-v3, rank 32), applied via PEFT adapter scaling at strength λ. Algorithm: genetic search — population 16, 10 generations, elitism (top 6 carried over). Each generation: every genome is generated best-of-8 and its fitness = the mean target-emotion strength over those 8 candidates (EmoNet score for the emotion, minus 0.25·max(0, WER−0.6) as a light intelligibility guard). Offspring = 60% crossover of two elites + mutation, 40% mutation of one elite. Genome (what evolves): the GENERAL voice description, the SCRIPT delivery cue, the carrier sentence, the LoRA strength λ (seeded 0.75–1.5, mutated within 0.4–2.0), and the sampling params (audio temperature, top_p, top_k). Mutation operators: 35% regenerate the description (new intensity + manner adjectives), 20% swap the carrier sentence, 23% jitter λ by ±0.25/0.5, 12% jitter temperature ±0.1, 10% change top_k. LoRA on vs off: the two LoRA OFF cards are baselines generated with the adapter disabled (λ=0) — a plain neutral prompt, and the best steering prompt without the LoRA. Every LoRA ON hall-of-fame recipe below was generated with the LoRA active at the evolved λ shown on the card. Seed is derived from the emotion name (deterministic).
Baselines — LoRA disabled
NEUTRAL promptLoRA OFF (λ=0)
mean-emo 0.108 · best 0.19 · blend 0.21 · genu 0.05 · WER 0.47
STEER promptLoRA OFF (λ=0)
mean-emo 0.019 · best 0.13 · blend 0.68 · genu 0.21 · WER 0.12
GENERAL: A voice unmistakably expressing concentration, a focused, deliberate voice, fully concentrated, measured and precise, pouring out.
SCRIPT: (intensely focused and concentrated) "I need to stop by the store and pick up a few things on the way home."
Hall of fame — top 7 evolved recipes (all generated WITH the emotion LoRA active)
#1LoRA ON · λ=2.0 (200%)
mean-emo 0.217 · best 0.33 · blend 0.37 · genu 0.10 · WER 0.28
GENERAL: A voice intensely expressing concentration, a focused, deliberate voice, fully concentrated, measured and precise, raw and unfiltered.
SCRIPT: (intensely focused and concentrated) "The report is on the desk, right next to the printer."
mean-emo 0.207 · best 0.27 · blend 0.21 · genu 0.10 · WER 0.35
GENERAL: A voice powerfully expressing concentration, a focused, deliberate voice, fully concentrated, measured and precise, pouring out.
SCRIPT: (intensely focused and concentrated) "The report is on the desk, right next to the printer."
mean-emo 0.202 · best 0.30 · blend 0.29 · genu 0.08 · WER 0.00
GENERAL: A voice to the extreme expressing concentration, a focused, deliberate voice, fully concentrated, measured and precise, impossible to hide.
SCRIPT: (intensely focused and concentrated) "The report is on the desk, right next to the printer."
mean-emo 0.200 · best 0.33 · blend 0.31 · genu 0.15 · WER 0.43
GENERAL: A voice to the extreme expressing concentration, a focused, deliberate voice, fully concentrated, measured and precise, raw and unfiltered.
SCRIPT: (intensely focused and concentrated) "The report is on the desk, right next to the printer."
mean-emo 0.198 · best 0.29 · blend 0.29 · genu 0.12 · WER 0.38
GENERAL: A voice deeply expressing concentration, a focused, deliberate voice, fully concentrated, measured and precise, building and building.
SCRIPT: (intensely focused and concentrated) "The report is on the desk, right next to the printer."
mean-emo 0.192 · best 0.32 · blend 0.33 · genu 0.17 · WER 0.57
GENERAL: A voice powerfully expressing concentration, a focused, deliberate voice, fully concentrated, measured and precise, building and building.
SCRIPT: (intensely focused and concentrated) "The report is on the desk, right next to the printer."
mean-emo 0.191 · best 0.35 · blend 0.25 · genu 0.08 · WER 0.56
GENERAL: A voice intensely expressing concentration, a focused, deliberate voice, fully concentrated, measured and precise, raw and unfiltered.
SCRIPT: (intensely focused and concentrated) "The report is on the desk, right next to the printer."
