R/BOBSERVATORY
CONNECTING
SPECIMEN 001RATTUS NORVEGICUS · SYNTHETIC

A SMALL NERVOUS SYSTEM. A VERY LARGE WORLD.

RATBRAIN.

A simulated rat, loose on the web.
Spiking. Exploring. Learning slowly.

CA: TBA

Enter the observation room
NEURAL SPECIMEN / 01

01 / OBSERVATION ROOM

A life in progress.

An autonomous, ongoing simulation. You're here to observe.

READ-ONLY OBSERVATORYFollow the evidence
WAITING FOR CONNECTION/PUBLIC WEB
NO SAMPLES YET

The room is ready.

Connecting to the autonomous session.
ACTUAL BROWSER FEED
MODEL INPUT: LINK BOXES, NOT PIXELS
CURRENT ACTION
X — / Y —HEADING —
AWAITING MODEL
SESSION UPTIME00:00
MOTOR DECISIONS
PAGES THIS SESSION
LINKS FOLLOWED
SUCCESSFUL SCROLLS
RECOVERIES

A motor decision may turn or rest; it does not necessarily open a page. Recoveries are automatic navigation resets.

02 / BRAIN & BEHAVIOR

Every movement begins here.

Persistent neurons. Real spikes. A decision you can trace—and see which channels shaped it.

NEURAL TIME
INTEGRATION
COMPUTE / WINDOW
NEURAL : WALL TIME

A nervous system in motion

ATLAS SHAPE / SYNTHETIC UNITS
RECORDED SPIKES · SLOWED FOR OBSERVATION
DRAG TO ORBIT · SCROLL TO MAGNIFY
— NEURONS— SYNAPSES— ABS RECURRENT / EXTERNAL DRIVE

Waiting for the live neuron sample. Rat anatomy: Waxholm Space atlas. Neuron positions illustrate functional populations; they are not measured cell locations.

The last neural moment

— ms
0 msTIME → / 256 SAMPLED NEURONS
PLACE · 64 CHANNELSGRID · 36HEAD DIRECTION · 12SENSORY · 16
MEMBRANE
SPIKES / WINDOW
THRESHOLD

Inspection changes only your view. The simulation continues independently.

01 · OBSERVATIONAwaiting a pageDOM link geometry + luminance and contrast
02 · NEURAL STATEAwaiting spikesVoltage → threshold → filtered spikes
03 · POLICYAwaiting a decisionSampled from the learned distribution
04 · BODY & BROWSERAwaiting an outcomeMovement and navigation are recorded

Before the next move

ACTION PROBABILITY

The next decision will appear here.

VALUE · RECENT UPDATES
TD ERROR · RECENT UPDATES
MODEL FOLLOWS / 100 · SESSION
LEARNING UPDATES · LIFETIME
Latest Σ|Δw|Reward received

TD updates change the policy weights. The follow-rate curve is session-only, calculated over up to the latest 100 decisions; it is an outcome measure, not proof that learning improved navigation.

Field notes

LIVE EVENT RECORD

No observations recorded yet.

MODEL DECISIONBROWSER ACTIONSUPERVISOR

LINK SELECTION · HEURISTIC · NON-NEURAL

Heuristic link selection

CURRENT PAGE

Awaiting a recorded link decision.

This hand-coded rule ranks allowed DOM links using visit counts, page position, label length, and distance, then makes a reproducible draw. It is outside the neural model and does not read or understand the excerpt.

THE WORLD IT REMEMBERS

Places leave a trace.

— KNOWN PAGES

Its first encounters will appear here.

This is a persistent SQLite visit-count table that changes link scores. It stores page URLs, titles, timestamps, and counts; it does not store or understand page content.

03 / THE RECORD

Don't take our word for it.

Inspect the live decision record, current chain head, and public methods.

Download terminal hashes

Read methods & evidence Download behavior baselines SOURCE PRIVATE DURING DEVELOPMENT · PUBLIC GITHUB PLANNED

SELF-ATTESTED RECORD

The chain head and terminal-hash ledger are published by the same operator. They can expose later changes to a saved record, but they do not independently prove when the run occurred.

RECORDING SEGMENTAwaiting first record

— recorded decisions

SHA-256 CHAIN HEAD

Each decision commits to the record before it.

Inspect the latest live decision record
Loading the recorded observation, neural output, and action outcome…
What the current record establishes

The live record binds each saved observation, neural output, chosen action, and browser outcome to the prior decision with SHA-256. The public terminal-hash ledger preserves each segment's final head.

This is operator-published evidence, not independent attestation or biological validation. Source release and public replay tooling are planned for a later GitHub publication.

04 / THE FOUNDATION

A nervous system, made observable.

Read the science

RATBRAIN brings rat-inspired spatial encodings, a persistent spiking circuit, and online weight updates into one observable simulated animal. Filtered spikes feed a policy that drives real browser actions. Its state survives the next decision—and the next session.

I

A brain with continuity

Membrane potentials evolve. Synaptic currents decay. Neurons fire and recover. Spatial tuning and sensory cues shape the 128 filtered channels that enter the policy.

Inspect the circuit ↑
II

Experience changes its weights

New destinations produce reward; TD updates change action and value weights. The session outcome curve shows whether model-selected follows rise rather than assuming improvement.

Inspect the updates ↑
III

An experiment with a record

Inputs, spikes, choices, and outcomes are recorded together. Inspect the live record and its self-attested terminal hashes; public source and replay tooling are planned.

Examine the record ↑
MODEL SCOPE

A reduced, rat-inspired spiking circuit—not a whole-brain reconstruction. Neuron counts and physical units describe the simulation; anatomy comes from a rat atlas; placement of model units is illustrative. The methods specify each assumption.