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

krit

-0.83

 Chaser

-0.73

livion

-0.71

endar

-0.70

elist

-0.69

Artist

-0.67

eq

-0.66

 understatement

-0.65

gression

-0.65

artist

-0.64

POSITIVE LOGITS

 Vanguard

0.68

 Blend

0.65

 Reverse

0.65

eat

0.64

 Running

0.63

ovo

0.63

 Runner

0.62

 Personally

0.62

 Ways

0.61

aults

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.