10-generation evolution of prompt + LoRA-strength + sampling to maximize Emotional_Numbness. Fitness = mean over 8 candidates. Best evolved 0.598 vs prompt-without-LoRA 0.008 (LoRA lift +0.589) vs neutral 0.054; winning λ≈0.75.
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 0.054 · best 0.12 · blend 0.21 · genu 0.05 · WER 0.47
STEER promptLoRA OFF (λ=0)
mean-emo 0.008 · best 0.02 · blend 0.38 · genu 0.12 · WER 0.39
GENERAL: A voice overwhelmingly expressing emotional numbness, a flat, hollow, emotionally numb voice, detached and empty, drained of all feeling, with every word.
SCRIPT: (flatly, hollow and emotionally numb) "The report is on the desk, right next to the printer."
Hall of fame — top 7 evolved recipes (all generated WITH the emotion LoRA active)
#1LoRA ON · λ=1.5 (150%)
mean-emo 0.598 · best 0.68 · blend 0.43 · genu 0.12 · WER 0.98
GENERAL: A voice utterly expressing emotional numbness, a flat, hollow, emotionally numb voice, detached and empty, drained of all feeling, in every breath.
SCRIPT: (flatly, hollow and emotionally numb) "I feel it, all of it, right now."
mean-emo 0.507 · best 0.66 · blend 0.46 · genu 0.12 · WER 0.75
GENERAL: A voice powerfully expressing emotional numbness, a flat, hollow, emotionally numb voice, detached and empty, drained of all feeling, with every word.
SCRIPT: (flatly, hollow and emotionally numb) "I feel it, all of it, right now."
mean-emo 0.551 · best 0.67 · blend 0.40 · genu 0.13 · WER 1.00
GENERAL: A voice powerfully expressing emotional numbness, a flat, hollow, emotionally numb voice, detached and empty, drained of all feeling, in every breath.
SCRIPT: (flatly, hollow and emotionally numb) "I feel it, all of it, right now."
mean-emo 0.447 · best 0.64 · blend 0.43 · genu 0.26 · WER 0.91
GENERAL: A voice powerfully expressing emotional numbness, a flat, hollow, emotionally numb voice, detached and empty, drained of all feeling, with every word.
SCRIPT: (flatly, hollow and emotionally numb) "I feel it, all of it, right now."
mean-emo 0.413 · best 0.66 · blend 0.31 · genu 0.15 · WER 0.77
GENERAL: A voice powerfully expressing emotional numbness, a flat, hollow, emotionally numb voice, detached and empty, drained of all feeling, in every breath.
SCRIPT: (flatly, hollow and emotionally numb) "The train leaves from platform four every hour on the hour."
mean-emo 0.451 · best 0.66 · blend 0.46 · genu 0.13 · WER 0.96
GENERAL: A voice to the extreme expressing emotional numbness, a flat, hollow, emotionally numb voice, detached and empty, drained of all feeling, impossible to hide.
SCRIPT: (flatly, hollow and emotionally numb) "You have no idea what this means."
mean-emo 0.406 · best 0.66 · blend 0.49 · genu 0.13 · WER 0.81
GENERAL: A voice viscerally expressing emotional numbness, a flat, hollow, emotionally numb voice, detached and empty, drained of all feeling, raw and unfiltered.
SCRIPT: (flatly, hollow and emotionally numb) "I feel it, all of it, right now."
