INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

 Detected

-0.81

ãĥİ

-0.66

andom

-0.65

Rated

-0.64

brance

-0.63

borgh

-0.63

advertising

-0.62

 Honour

-0.61

 externalToEVAOnly

-0.61

PASS

-0.61

POSITIVE LOGITS

trak

0.73

swick

0.68

tub

0.68

fficiency

0.67

yk

0.66

 glim

0.65

lag

0.65

itness

0.64

 blast

0.63

iens

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.