INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

lift

-0.72

 overnight

-0.70

Enlarge

-0.67

header

-0.67

 Allaah

-0.66

inline

-0.65

 Leilan

-0.64

Fly

-0.64

Header

-0.62

Ri

-0.61

POSITIVE LOGITS

 pacing

0.79

 throats

0.76

iosyn

0.70

chest

0.69

 mouths

0.66

senal

0.66

iaries

0.64

 pens

0.62

idth

0.61

 Finch

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.