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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Tens

-0.68

 alienation

-0.63

forced

-0.61

xp

-0.61

 Whites

-0.61

venge

-0.60

Uri

-0.60

 disbelief

-0.60

gen

-0.59

 residence

-0.57

POSITIVE LOGITS

nir

0.84

itzer

0.77

olesterol

0.75

 Carbuncle

0.71

ancer

0.71

 Dragonbound

0.71

hips

0.71

£ı

0.70

ois

0.67

Aust

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.