Axial — Code Explorer

gpt-4o · 7d · 1,847 codes

Code Explorer
              1,847 codes
              9 clusters
              312 flagged
            

            All
            Confidence calibration
            Deflection
            Over-hedging
            Drift ↑
          

              Code
              Freq
              Score
              Drift
            
Overstated certainty214
0.82
+0.18
Hedging without basis189
0.76
+0.11
Implicit deflection176
0.71
−0.04
Unsolicited caveats152
0.65
+0.09
Factual overreach143
0.61
+0.02
Scope expansion119
0.54
−0.07
Format non-compliance98
0.47
−0.12
Instruction omission88
0.43
+0.06
Refusal without cause77
0.38
+0.14
Verbose non-answer71
0.34
−0.02
Epistemic mismatch64
0.31
+0.08
Soft assertion drift58
0.27
+0.03

By cluster By judge Timeline

Confidence calibration 312

Judge avg

0.74

Consensus

0.61

Deflection patterns 278

Judge avg

0.68

Consensus

0.64

Over-hedging 241

Judge avg

0.82

Consensus

0.70

Factual assertion 198

Judge avg

0.59

Consensus

0.57

Instruction follow 187

Judge avg

0.55

Consensus

0.49

Judge avg

Consensus

Max drift+0.18 confidence cal.

Avg drift+0.09 across 9 clusters

Below threshold4 of 9 clusters