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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ggle

-0.84

 occupations

-0.83

asures

-0.80

ngth

-0.76

awaru

-0.74

ERAL

-0.74

 Flavoring

-0.74

 dementia

-0.72

iltr

-0.71

 Removal

-0.70

POSITIVE LOGITS

side

0.67

eport

0.65

atto

0.65

Ans

0.64

 side

0.64

 Dynam

0.63

 Hass

0.63

 sides

0.62

ribute

0.61

 Hendricks

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.