INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

tf

-0.74

py

-0.70

Lenin

-0.69

Self

-0.63

 firsthand

-0.62

 proceeds

-0.62

encing

-0.62

iste

-0.61

Already

-0.61

boxing

-0.60

POSITIVE LOGITS

eatures

0.79

 Hawk

0.72

actionDate

0.68

oun

0.64

 Magnum

0.61

 Tempest

0.60

unct

0.60

Naj

0.60

Ori

0.60

 Accuracy

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.