INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

âĢİ

-0.72

aq

-0.68

abad

-0.65

oward

-0.64

izoph

-0.61

iona

-0.61

 avalanche

-0.60

JA

-0.60

 specialization

-0.59

 shortage

-0.58

POSITIVE LOGITS

heres

0.67

roman

0.67

uploads

0.64

anus

0.63

Anonymous

0.62

hidden

0.61

interstitial

0.61

soever

0.60

hered

0.60

INTER

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.