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

LAB

-0.76

FontSize

-0.76

é¾

-0.73

 sideline

-0.72

Market

-0.71

 Crus

-0.70

Wars

-0.67

æ©Ł

-0.66

 Ulster

-0.66

 Beir

-0.65

POSITIVE LOGITS

ifice

0.71

ials

0.69

agonist

0.68

irection

0.68

ocent

0.66

ozo

0.66

anced

0.65

unction

0.63

ox

0.63

ost

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.