deepseek-v4.1-flash @ off · dsh
dshoff
- config
- deepseek-v4.1-flash-eff-off--cli--dsh--dev
- batch
- 2026-07-19--unified
- transport
- dsh-0.1.0-rc.8-headless-isolated-home
- protocol
- cli · api
- served model
- deepseek-flash
- captured
- Jul 18, 2026, 11:54 AM · N=2
live
no screenshot for this slot
The screenshot is the exact frozen pixels the similarity channels measured. Run live re-executes the raw output in a sandboxed iframe; its rendering depends on your browser and fonts and is not guaranteed to match the measured pixels.
similarity neighbors
nearest
claude-opus-5-5 @high · CC0.66claude-opus-5-thinking @max · cursor0.66claude-opus-5 @medium · cursor0.65claude-opus-5 @low · cursor-xl0.65claude-opus-5-thinking @low · cursor-xl0.65farthest
qwen3.6-plus · go0.50grok-4.3 @low · api0.52grok-4.3 @medium · api0.52grok-4.3 · api0.52deepseek-v4-pro-0812 @none · api0.53within session 2026-07-19--unified
similarity measures convergence within this session — not quality.
No valid output
Every slot for this config failed to render — a broken output is data, not a hidden gap. The state chips and send-log below show exactly what happened.
Fidelity — does the card show the right data?
No fidelity verdict — this slot is — (fidelity is computed for valid slots only).
Source — raw model output, one byte unchanged
No card.html for this slot (—).
L1 — cheap deterministic descriptors
No L1 — this slot is —.
This slot's telemetry
config.json596 B
{
"N": 2,
"auth": {
"credential_ref": "WCB_DEEPSEEK_KEY",
"method": "api-key"
},
"billing": "api",
"config_id": "deepseek-v4.1-flash-eff-off--cli--dsh--dev",
"effort": "off",
"family": "deepseek",
"m": {
"min": 2,
"q": 2
},
"model_id": "deepseek-v4.1-flash",
"protocol": "cli",
"served_model": "deepseek-flash",
"telemetry": {
"completion_tokens": 118278,
"cost_usd": null,
"prompt_tokens": 110274,
"total_tokens": 228552,
"wall_ms": 1211348
},
"transport": "dsh-0.1.0-rc.8-headless-isolated-home",
"vendor_sanction_ref": null
}send-lognot on disk
No send-log on disk for this config.