INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.0.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

rily

-0.85

Interstitial

-0.78

oro

-0.76

*/

-0.74

Export

-0.69

 Gujar

-0.69

-=

-0.67

GPL

-0.66

Ptr

-0.65

 Expend

-0.65

POSITIVE LOGITS

hol

1.02

engineering

0.75

ovie

0.67

wire

0.67

utterstock

0.66

dream

0.66

urdue

0.65

hospital

0.63

vision

0.61

draw

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.