INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

çĦ

-0.87

sbm

-0.76

åŃ

-0.76

Occ

-0.73

ecause

-0.69

pta

-0.66

AX

-0.65

amazon

-0.65

Japanese

-0.64

Accessory

-0.64

POSITIVE LOGITS

ascript

0.78

enegger

0.77

burgh

0.74

etsk

0.71

 firsthand

0.70

 destro

0.69

mington

0.68

apult

0.68

ilver

0.67

opian

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.