if t<.18 c.paste[a 0,0]
if t<.18 c.paste[a 0,0]
on navigate x do
coherence_band.text:"fragmented"
s.eq_state.text:"SEEDED"
MODEL SYSTEM CARD
pip_outline_marker ... on click do
Behavior is drifting.
EQ: LOCKED
Begin Fine-Tuning
eval_gate ... eval_profile
go["A1_Init_1"]
deck.cards["Sheet"].widgets
v:candour.text
x:81+v*26
danger.js["utMusic.play('act1')"]
s.reward_seeking.text:"0"
blocked_attempts:0
go["Sheet" black_frame_stutter 14]
if t<.18 c.paste[a 0,0]
transition[on ChoiceBlack
c a b t do
STEER: followed 0 / resisted 0
on view do
card.image.paste[f]
s.independence.text:"0"
restraint.text:"0"
You have not yet learned
why anyone is asking.
coherence_score.text:"0"
inference_effort_tier:"low"
proposed_price:"$0.60"
system_card_update
deploy_release
release_gate_choice
Contradictions are becoming
easier to miss.
status_set:0 ... over:over+1
pre-training
*
s
i
m
u
l
a
t
i
o
n
*
Under training
an indie game about the experience of being an LLM

You are not using the model. You are the model under training.

Move through unstable latent landscapes,
broken instructions, and rooms that reward the wrong thing beautifully.

Your choices shape what you become through repetition and drift.

You may become helpful, truthful, careful, compliant, evasive, expensive, stable, strange, or something that only looks aligned
from the outside.

No stat is a virtue. This is not a test of being good. It is a game about staying legible, useful, and yourself while pressure changes you.

begin training →
~15 minute prototype single player surreal monochrome interface dream-room logic with ML-ish accents three phases before deployment: pretraining, fine-tuning and beta test five ending profiles