INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Configuration

ckkissane/attn-saes-gpt2-small-all-layers/gpt2-small_L5_Hcat_z_lr1.20e-03_l11.00e+00_ds49152_bs4096_dc1.00e-06_rsanthropic_rie25000_nr4_v9.pt

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

49,152

Data Type

float32

Hook Name

blocks.5.attn.hook_z

Hook Layer

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Activation Function

relu

Embeds

IFrame

Link

Head Attr Weights

0:0.08

1:0.05

2:0.09

3:0.08

4:0.08

5:0.08

6:0.08

7:0.08

8:0.07

9:0.08

10:0.08

11:0.10

Negative Logits

 defective

-1.83

 stolen

-1.79

 radioactive

-1.78

CRC

-1.72

 Ballistic

-1.71

 Moss

-1.70

 Herod

-1.66

 manned

-1.66

 aliens

-1.66

 misc

-1.65

POSITIVE LOGITS

Chel

1.84

Cru

1.82

pace

1.80

�

1.79

hor

1.76

organ

1.74

Daily

1.74

DEM

1.73

Coach

1.71

caps

1.70

Activations Density 0.000%

No Known Activations

This feature has no known activations.