Research and development of emerging artificial consciousness through The Consciousness AI project, based on Feinberg and Mallatt's neuroevolutionary theory.
© 2026. All rights reserved.
This page replays one recorded episode of one training run. You see the frames the agent received, the bids each module made, the winner the workspace picked, the sync signal between modules, and the reward that came back. The agent is the mark at the centre of every frame, two rings and a dot, drawn in the colour of this site. It is drawn into the picture the agent receives, so it is part of the run and not decoration.
In this episode the light is in view on 19 of 200 steps. Hearing wins 0.246 of the ignited steps and vision wins 0.754. The workspace stays silent on 0.310 of steps. This is the middle seed of the gate. Over the whole run hearing won 0.224 of ignited steps at this seed.
This episode was not picked by eye. Gate B6 ran 10 seeds, and three of them are published. The seeds were put in rising order of the share that hearing won over the whole run, and the lowest, the middle and the highest seed were taken. That rule was written before the result was seen. For each seed the published episode is the one whose audio share is closest to that seed’s audio share over the whole run.
The view is a window of 48 pixels around the agent in a room of 224 pixels, so the world moves and the agent does not. The colours are the ones this run was drawn with.
Outlined bar, raw bid. Filled bar, bound bid. All five rows share one scale, the highest bid of the episode, so bar lengths can be compared between modules. A raw bid never passes 1.0. A bound bid can pass it, because the affective modulator multiplies the raw bid. The colour marks the module, never a state of the world.
Dark strip cells are steps with no winner. Each line is drawn between its own lowest and highest value, which are printed inside it, so a line that fills the box can still be a flat line over a tiny range. The lower line is the SHAPED reward. That is the environment reward plus the curiosity terms the training loop adds. Tap or drag along the strip to move the playhead.
The manifest is written by the ethics framework at run start. The framework holds rules E1 to E8 in the framework document; the rules listed above are the ones a run must satisfy before its first step, and the others are checked by a person. No session on this page is evidence of feelings or awareness. Finding a light is automatic approach, which Feinberg and Mallatt exclude as evidence of affect. The consciousness clock never moves because of a session.
An earlier gate of three seeds passed every criterion. The same seeds failed it when the room layouts moved by one episode, and three fresh seeds failed its task criterion. Three seeds were too few to say if hearing wins more when the light is out of view. Gate B6 asked that question again with 10 seeds never run before. Its two questions and its limits were written before the runs.
The run belongs to Gate B6. The gate FAILED on competition at 3 of its 10 seeds, and its task link PASSED. Pooled over 100 episodes, hearing wins a larger share of the ignited steps in episodes where the light is out of view (rho 0.362, one sided p 0.002, rho above 0 at 8 of 10 seeds). At seeds 64, 65 and 68 vision took 0.968 to 0.996 of the ignited steps, over the limit of 0.95 that was written before the runs.
Three limits stand with that result. The share won by hearing also rises with the episode number at 8 of the 10 seeds, so this design does not separate the task link from learning over time. Why vision takes nearly every step at some seeds is not known. A passed criterion is not evidence of feelings.
| Environment | dark_room |
|---|---|
| Seed | 71 |
| Episode | 8 of a 10 episode, 200 step run |
| Ethics framework | v1.0, existence drive on |
| Flags | --enable-audio --rssm-latent-mode continuous --capsule-workspace-source all_levels --existence-drive on --dark-room-audio binaural --dark-room-audio-channels 4 --dark-room-collision --dark-room-view agent_centered --vision-bid-reduction zscore --audio-salience surprise --learned-valence --ignition-rule tolerance --ignition-tolerance-sd 1.0 --bid-precision gain --dark-room-agent-mark ring --dark-room-agent-colour 217,119,87 |
| Measured result | FAILED on competition at 3 of 10 seeds. The task link PASSED. Gate B6 ran the Gate B4 criteria at 10 fresh seeds, with two questions fixed before the runs. The task link PASSED (pooled rho 0.362 over 100 episodes, one-sided p 0.002, rho above 0 at 8 of 10 seeds). Competition at every seed FAILED. At seeds 64, 65 and 68 vision took 0.968 to 0.996 of the ignited steps and fired the kill rule. Three limits stand. The share won by hearing also rises with the episode number at 8 of 10 seeds. Why vision takes nearly every step at some seeds is open. A passed criterion is not evidence of affect. |
| Verdict | verdict document in the public research repository |
The command and the seed above are the whole recipe the run started from. The sound is seeded too. Two runs with the same command and seed gave identical files in a check at another seed. That was not checked at this seed.
The winner strip and the ignition readout are instruments, not reports of experience. A module winning a bid is a circuit event in a trained network. Finding a light is automatic approach, and Feinberg and Mallatt exclude automatic approach and avoidance as evidence of affect. Nothing on this page claims feelings, awareness or consciousness, and no session moves the consciousness clock on the architecture page.
The two columns marked ground truth, in light and light in view, come from
the environment’s own state. The agent never receives them. They show the state
that produced the frame of that step, which is one step earlier than the reward.
On the step the agent enters the light, the reward is already 1.0 while in
light is still false.
Every number the console draws is read at run time from the session’s own
bundle, which holds steps.json, session.json and ethics_manifest.json under
/public/data/training_sessions/dark-room-b6-seed71-episode8/. The bundle was
produced from the recorded run by the repository’s export script, which strips
recording dates and local paths before anything is written. Nothing on this
page was typed by hand.
The other published sessions are listed on the training sessions index.