CITADEL
StandbyNothing is live right now.
Eval accuracylast measured Wed 18 Nov, 00:14 UTC
%
0.8880444 of 500 eval, 56 wrong
▲ +72.00 pts across 4 evals▲ +3.60 pts on the eval before
0%25%50%75%100%0%25%50%75%100%0%25%50%75%100%

Familiar decider: ModernBERT-base (backfilled)

finishedRadeon 8060S

One of the small models that decide whether Familiar, a desktop companion, reacts to something happening on the PC, and how. Backfilled: recorded after the fact from the run's own logs and records. Times inside a run are estimated, and there are no GPU sensor readings.

started
Tue 17 Nov, 23:55 UTC
ran for
14:42
config
learning rate 0.00005, batch 16, seed 0 of 3

run3/5step1125/1125epoch3/3config3/3seed0/3lr0.00005batch16modelModernBERT-base 149Mrev8949b90items6000+500

Eval measures

Elapsed

294.0 s

epoch interval

294.0 sto eval 2294.0 sto eval 3294.0 sto eval 4

294.0 s to eval 2, 294.0 s to eval 3, 294.0 s to eval 4

1125 of 1125 steps, 3 of 3 epochs, 100%

0.78 s per step, 76.5 steps a minute

Eval accuracy by eval

+0.0360

▲ on the eval before

eval 1 0.1680, eval 4 0.8880

Against the other runs

3 of 5

by last eval accuracy

best so far 0.9180, run 5

▲ +0.0240 against run 2

Set-up

6000 train, 500 eval

synthetic, learning rate 0.00005, batch size 16, epochs 3, max length 512, precision bf16-mixed, loss ce, optimizer adamw, scheduler linear, seed 0

Seed spread, observed

0.0300
0.850.95

3 finished seeds

Input

  • Runs128 runs
  • Agentsnone showing
  • Treesnone showing
  • Projectsnone showing

Check

Every entry is checked on its way out; some are held for review.

Not shown on any run: data samples, predictions, raw logs, file paths, machine names

Delay

2 min, set

Watching the agents work, about 2 min behind.

One way

Nothing here reaches the agents.

This page

This page updates by itself.

visits counted with Google Analytics · choose

Runsheight is eval accuracyall recordings ▸ | finished× failed■ stopped○ lost● live

3 of 5 finished, 0.8880, config 3 of 3 (learning rate 0.00005, batch 16), seed 0 of 3, Tue 17 Nov, 23:55 UTC

Runs as a table
StartedStatusEval accuracyStepsConfig
Wed 18 Nov, 02:41 UTC| finished0.91801125 of 1125learning rate 0.00005, batch 16, seed 2 of 3
Wed 18 Nov, 02:22 UTC| finished0.89401125 of 1125learning rate 0.00005, batch 16, seed 1 of 3
Tue 17 Nov, 23:55 UTC| finished0.88801125 of 1125learning rate 0.00005, batch 16, seed 0 of 3
Tue 17 Nov, 23:36 UTC| finished0.86401125 of 1125learning rate 0.00002, batch 16, seed 0
Tue 17 Nov, 23:17 UTC| finished0.85801125 of 1125learning rate 0.0001, batch 16, seed 0

Plate

Model
answerdotai/ModernBERT-base, 149M parameters revision 8949b90
Data
Familiar decisions, shared draw, synthetic6000 train500 eval
Config
learning rate 0.00005, batch size 16, epochs 3, max length 512, precision bf16-mixed, loss ce, optimizer adamw, scheduler linear, seed 0
Hardware
Radeon 8060S, ROCm

Log

Since startMark
0.0 sphase starting
0.0 seval 1
294.0 seval 2
294.0 sepoch 1
588.0 seval 3
588.0 sepoch 2
882.0 seval 4
882.0 sepoch 3
12 h 56 min laterwent public

Seeds

SeedStatusEval accuracy
seed 0| finished0.8880
seed 1| finished0.8940
seed 2| finished0.9180

observed spread 0.0300, 3 finished seeds

Readings

StepSecondsEval accuracy
00.00.1680
375294.00.8260
750588.00.8520
1125882.00.8880

Worked out

ReadoutValueWorked out from
Run complete100%1125 steps over 1125
Step time0.78 s294.0 s between epoch marks, over 375 steps an epoch. The reading interval holds the eval, so it is left out
Steps per minute76.560 s over the step time of 0.78 s
Epoch interval294.0 sMean of 2 gaps between epoch marks
Eval accuracy, as shown88.8%0.8880 times 100, to one decimal. Changes on that plate are percentage points
Eval accuracy since first eval+0.72000.8880 minus 0.1680
Eval change+0.0360Eval 4 at 0.8880 minus eval 3 at 0.8520
Wrong answers56500 eval items minus 444 correct. 444 over 500 is the only count that gives 0.8880
Rank3 of 5By last eval accuracy, across this project's 5 runs with an eval
Best so far0.9180Highest last eval accuracy, run 5
Against the run before+0.02400.8880 minus 0.8640 on run 2
Seed spread, observed0.0300High 0.9180 minus low 0.8880 across 3 finished seeds
Items per second20.41 step over 0.78 s, times batch size 16