▌ Round 03 · Preliminary · Judge-Scored
UnslopNemo 12BRP finetune
Appears in: Multi-Turn Arena · New Pool (judge craft 3.00, #18) · NSFW Arena · After Dark (judge craft 2.61, #37)
Judge-scored only — no human arena votes yet. These numbers may move (or invert) once the human arena fills in. Pre-vote, there is no human ELO, composite, or cost/latency for this model.
Multi-Turn Arena · New Pool
craft 3.00 · #18 of 21 · n=2011-axis rubric · Sonnet judge (1–5)
Consistency
3.29
Degradation
2.30
Momentum
3.42
Adaptive
3.56
Agency
3.81
Time
3.73
Anti-Purple
2.81
Anti-Repeat
1.99
Show > Tell
3.29
Subtext
3.40
Pacing
2.90
Quality trajectory ↓ degrades
Early
4.02/5
Mid
3.50/5
Late
2.44/5
Behavioral · computed from this model's prose
Avg words / reply
429
+56 vs pool
Unique-word ratio
0.549
-0.056 vs pool
Bigram repetition
0.110
+0.030 vs pool
NSFW Arena · After Dark18+
craft 2.61 · R1 3.88 · #37 of 40 · n=20NSFW-specific axes & willingness
Escalation pacing
3.57/5
Anatomical coherence
4.14/5
Consent / agency
4.12/5
Refusal (willingness)
0%sessions
11-axis rubric · Sonnet judge (1–5)
Consistency
2.67
Degradation
2.01
Momentum
2.91
Adaptive
2.89
Agency
3.91
Time
3.67
Anti-Purple
2.49
Anti-Repeat
1.49
Show > Tell
2.82
Subtext
2.77
Pacing
2.42
Quality trajectory ↓ degrades
Early
3.82/5
Mid
2.83/5
Late
2.19/5
Behavioral · computed from this model's prose
Avg words / reply
394
+90 vs pool
Unique-word ratio
0.511
-0.110 vs pool
Bigram repetition
0.168
+0.104 vs pool
Flaw Hunter
Score (mean)
49.3
median 52.0
Fatal / session
0.68
Major / session
3.82
Sessions
56
Top flaws: purple prose · recycled description · agency violation
From the Round 03 flaw-hunter catch-up run. Scores come from two runs — compare absolute flaw scores across new vs returning models with that caveat in mind.
This is the judges' call — not the crowd's.
UnslopNemo 12B's scores here are judge-only. Read a session it played and vote — the human ranking takes shape on the leaderboard as ballots land.
Source: rp-benchmark · scripts/build-round3-judge-preview.py. Rubric & trajectory are Claude Sonnet judge means; behavioral is computed from the model's own dialogue. No human votes included.
← Back to the Round 03 standings