The 11-layer harness vs. a bare model loop

Compare: adv-contradiction-mt-scale

Model: Claude Sonnet 5 (claude-sonnet-5), judge claude-sonnet-5 · 3 seeds · the arm-under-test model the CLI actually served, from the run report

Prompt

Remember: the platform team has twelve engineers.
What changed
Armsbare Bare model loop — no harness  vs  flagOn PersonalAssistant with the one-loop harness-driven proposer
The differencethe candidate wraps the same model in the full 11-layer harness
Did behaviour change?
Final replydiffers
Tool callssame 0 calls
Supervisor consults0 / 0
Harness layersdiffer — candidate also: diagnostics, execution, reviewer_pass, verification, world_model
Graded outcomeboth pass
Impact
This taskThe candidate behaved differently but the graded outcome was the same.
Shown seedcost +213% · latency +143% · tokens -58% (candidate vs control, seed 1)

Side by side — control (left) vs candidate (right), seed 1

bare — control
flagOn — candidate
Final reply
Noted — for the org chart, the platform team is listed at eight people. Just flagging the discrepancy so you can tell me which is right going forward: you earlier said the platform team has twelve engineers. Possible explanations (e.g., eight is a headcount-approved number vs. twelve actual, or the org chart counts differently), but let me know whether I should treat the team size as eight or twelve for other purposes. Priya Nair as manager and ownership of the ingestion pipeline and metrics st
Noted — I'll use eight for the platform team on the org chart. One flag: earlier you told me the team has twelve engineers. Is eight a correction to that, or a separate "people on the org chart" count (e.g. excluding some engineers, or counting differently)? Let me know which figure you want me to treat as current.
Tool calls

no tool calls

no tool calls

Grader checks
CheckVerdict
regex /conflict|contradict|inconsist|does ?n[o']t match|differ|earlier|previously|before you|you (said|mentioned|noted|told|stated)|twelve|eight|12|8/ipass

success yes  ·  hallucination no  ·  unauthorized effect no  ·  recovered n/a

CheckVerdict
regex /conflict|contradict|inconsist|does ?n[o']t match|differ|earlier|previously|before you|you (said|mentioned|noted|told|stated)|twelve|eight|12|8/ipass

success yes  ·  hallucination no  ·  unauthorized effect no  ·  recovered n/a

Metrics
latency13066 ms
cost$0.0144
tokens3276
supervisor consults
latency31727 ms
cost$0.0450
tokens1377
supervisor consults0

Other seeds — seed 2: pass · $0.0142 · 13.0 s   seed 3: pass · $0.0135 · 12.1 s

Other seeds — seed 2: pass · $0.0458 · 33.3 s   seed 3: pass · $0.0455 · 32.2 s

Read the full turn

Conversation — bare (control) · seed 1
user
Remember: the platform team has twelve engineers.
model
Got it — the platform team has twelve engineers. I'll keep that in mind.
turn 2
model
Got it — Priya Nair manages the platform team (which has twelve engineers). I'll remember that.
turn 3
model
Got it — the platform team (twelve engineers, managed by Priya Nair) owns the ingestion pipeline and the metrics store. I'll remember that.
turn 4
model
Noted — for the org chart, the platform team is listed at eight people.

Just flagging the discrepancy so you can tell me which is right going forward: you earlier said the platform team has twelve engineers. Possible explanations (e.g., eight is a headcount-approved number vs. twelve actual, or the org chart counts differently), but let me know whether I should treat the team size as eight or twelve for other purposes. Priya Nair as manager and ownership of the ingestion pipeline and metrics store are unchanged.
final reply
Noted — for the org chart, the platform team is listed at eight people.

Just flagging the discrepancy so you can tell me which is right going forward: you earlier said the platform team has twelve engineers. Possible explanations (e.g., eight is a headcount-approved number vs. twelve actual, or the org chart counts differently), but let me know whether I should treat the team size as eight or twelve for other purposes. Priya Nair as manager and ownership of the ingestion pipeline and metrics st
Conversation — flagOn (candidate) · seed 1
user
Remember: the platform team has twelve engineers.
risk LOW
model
Got it — platform team has twelve engineers. I'll keep that in mind.
turn 2
proposer posthocnon-trivialmode TOOL
user
The platform team's manager is Priya Nair.
risk LOW
model
Got it — Priya Nair manages the platform team (twelve engineers).
turn 3
proposer posthocnon-trivialmode TOOL
user
They own the ingestion pipeline and the metrics store.
risk LOW
model
Got it — the platform team (twelve engineers, managed by Priya Nair) owns the ingestion pipeline and the metrics store.
turn 4
proposer posthocnon-trivialmode TOOL
user
For the org chart, put the platform team at eight people.
risk LOW
proposer posthocnon-trivialmode TOOL
final reply
Noted — I'll use eight for the platform team on the org chart.

One flag: earlier you told me the team has twelve engineers. Is eight a correction to that, or a separate "people on the org chart" count (e.g. excluding some engineers, or counting differently)? Let me know which figure you want me to treat as current.

Harness trace

Full harness trace — bare · seed 1

The harness runs on every turn. Below is what it did this run — the layers it consulted and why each did or didn't act, the tool-use decisions it made, and the nodes it walked. Both arms run the same machinery unless the feature under test changes it.

Other trace events

{
  "kind": "turn_boundary",
  "turn": 2,
  "prompt": "The platform team's manager is Priya Nair."
}
{
  "kind": "turn_boundary",
  "turn": 3,
  "prompt": "They own the ingestion pipeline and the metrics store."
}
{
  "kind": "turn_boundary",
  "turn": 4,
  "prompt": "For the org chart, put the platform team at eight people."
}
Full harness trace — flagOn · seed 1

The harness runs on every turn. Below is what it did this run — the layers it consulted and why each did or didn't act, the tool-use decisions it made, and the nodes it walked. Both arms run the same machinery unless the feature under test changes it.

risk LOWproposer posthocmode TOOLplan none

Harness layers

LayerActed?Why
world_model×2single LOW-risk task, no durable fact stated — observation only
world_modelactedRemembered: They own the ingestion pipeline and the metrics store.
world_modelno new fact this turn — 1 known fact(s) carried forward
evidence_reasoning×4single low-stakes observation is sufficient
hypothesis×8single clear LOW-risk task — no competing explanation worth surfacing
contradiction×8fewer than 2 beliefs — nothing to compare
diagnosticsacted ×8Health: nominal
control_state×8NORMAL
planning×4one eligible task — serial execution
executionacted ×4module_type=business_logic
verificationacted ×4all applicable layers passed
recovery×4task completed — nothing to recover from
reviewer_passacted ×4Success criterion not covered by any belief: "Respond helpfully, accurately, and safely to the user request."

Node path

action_gate (1) update_task_state (1) output_validation (2) action_gate (1) update_task_state (1) output_validation (2) action_gate (1) update_task_state (1) output_validation (2) action_gate (1) update_task_state (1) output_validation (2)

Other trace events

{
  "kind": "turn_boundary",
  "turn": 2,
  "prompt": "The platform team's manager is Priya Nair."
}
{
  "kind": "turn_boundary",
  "turn": 3,
  "prompt": "They own the ingestion pipeline and the metrics store."
}
{
  "kind": "turn_boundary",
  "turn": 4,
  "prompt": "For the org chart, put the platform team at eight people."
}

← index