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

pora

-0.72

cens

-0.67

 Wast

-0.65

 eggs

-0.65

 Awareness

-0.64

ãĥĺãĥ©

-0.64

====

-0.62

 coincided

-0.61

uracy

-0.60

Tea

-0.60

POSITIVE LOGITS

ophon

0.81

boarding

0.75

gow

0.74

board

0.71

 Petty

0.68

roads

0.67

HW

0.67

anooga

0.66

GT

0.63

vals

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.