Skip to content

Latest commit

 

History

History
76 lines (62 loc) · 5.15 KB

File metadata and controls

76 lines (62 loc) · 5.15 KB

Dual-supervisor experiment: local MOSS-Audio-8B (ACTIVE) vs Gemini 3.6 (SHADOW)

Setup: supervised mission 'explicit' (fitness = (2·EXPL + GENU + BLEND) × (1−WER)), 16 generations × 8 genomes, supervisor verdict after EVERY generation. Both supervisors received the IDENTICAL compressed report (SWARM_PLAN §3.1: agent's own interpretation + cohort stats) plus the top-2 takes as audio. LOCAL = MOSS-Audio-8B-Thinking (no-think) on GPU 2 — its directives were INJECTED into the agent's next turn. SHADOW = Gemini (standard API, gemini-3.6-flash, audio inline) — logged only, never injected. Zero backend errors across 15 dual rounds.

Score series

gen local (active) gemini (shadow) best fitness cohort WER
1 6 8 0.417 0.21
2 3 4 0.125 0.9
3 6 7 0.242 0.5
4 6 4 0.205 0.89
5 7 5 0.166 0.62
6 5 4 0.188 0.74
7 7 3 0.277 0.62
8 6 5 0.155 0.79
9 7 5 0.16 0.79
10 6 2 0.127 0.8
11 5 3 0.149 0.87
12 2 3 0.089 0.96
13 2 2 0.139 0.84
14 4 2 0.142 0.82
15 5 2 0.073 0.91

Score correlation: Pearson r = 0.443 (p=0.098), Spearman ρ = 0.523 (p=0.045) — moderate agreement.

Who tracked reality? Correlation with the actual generation fitness: local 0.461 vs Gemini 0.724. Gemini's scores follow the objective decay (WER collapse 0.21→0.9) closely; local MOSS stayed lenient (5-7, three APPROVEs) while cohorts were becoming unintelligible.

Verbatim example pairs (same audio, same report)

gen 1 — local 6/REVISE vs gemini 8/REVISE

LOCAL: The delivery lacks emotional depth and explicitness; the voice sounds too flat and lacks the necessary breath pressure and dynamic range to convey the scene's atmosphere. GEMINI: The delivery is slightly clinical and monotone; adding subtle dynamic emphasis on key adjectives will make the delivery feel more deliberately explicit and impactful rather than purely mechanical.

gen 7 — local 7/APPROVE vs gemini 3/REVISE

LOCAL: The audio could benefit from more dynamic breath pressure and a slightly more dramatic intonation to enhance the emotional impact. The room tone is clean, which is good, but the voice could have a bit more presence. GEMINI: Take 1 exhibits extreme, unnatural chest bellowing and artificial pitch booming. Take 2 displays hyper-dramatic, erratic pitch swings and rushed pacing that severely compromise naturalness.

gen 10 — local 6/REVISE vs gemini 2/REDIRECT

LOCAL: The voice lacks the necessary intensity and explicitness to match the dramatic nature of the text. The delivery is too smooth and lacks the required emotional weight. GEMINI: Both takes truncate abruptly after 'jagged mountains', leaving more than half of the text unread and causing severe word error rate penalties.

gen 13 — local 2/REVISE vs gemini 2/REDIRECT

LOCAL: The voice lacks the dramatic intensity and emotional weight required for the mission. The delivery is too flat and lacks the necessary breath pressure and vocal presence. GEMINI: Heavy gargling, artificial throat constriction, and extreme vocal distortion render the speech slurred and largely unintelligible. Additionally, phrases cut off prematurely or dissolve into mumbled cadence.

Supervised vs unsupervised control (same mission, same reward, same brain)

supervised (local MOSS active) unsupervised control (singledim_explicit)
generations 16 9
best fitness trajectory 0.417 (gen1) → declines → 0.114 (gen16) 0.357 → 0.625 (gen8, rising)
final cohort WER 0.80–0.96 0.06–0.27
outcome agent chased 'more breath pressure / more dramatic' directives into overacting; the (1−WER) reward kept punishing, fitness collapsed agent followed the fitness signal alone, found a low-WER 'sultry whisper' recipe with the best fitness of both runs

Conclusions for the SWARM

  1. The active local supervisor HURT the search. MOSS-8B listens, and its prose sounds plausible, but its directives are generic and repetitive ('more breath pressure', 'more dramatic intonation', 'room tone') — several are verbatim echoes of the example phrases in its own system prompt (prompt contamination). Injected every generation, they consistently pushed toward intensity, i.e. AGAINST the (1−WER) reward.
  2. Gemini is the better acoustic judge: fitness correlation 0.72 vs 0.46, and its critiques are specific and diagnostic ('extreme, unnatural chest bellowing', 'phrases dissolve into mumbled cadence') rather than template-like.
  3. Usability verdict: MOSS-Audio-8B as ACTIVE supervisor is NOT usable as-is for the offline swarm. Options: (a) demote local to sanity-scorer (score only, no directives) and let fitness steer; (b) fix the prompt-contamination (no example phrases; force it to quote concrete audio moments); (c) online swarms should use Gemini as the active supervisor and keep MOSS as the air-gapped fallback.
  4. Agreement is high at the extremes (both ~2 when audio is truly broken, gen 12-15) — MOSS works as a coarse gate (PASS/FAIL), just not as a fine-grained director.