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

Missing

-0.77

ENCY

-0.76

Bra

-0.76

BAT

-0.74

 Bland

-0.73

SHA

-0.71

Score

-0.68

Disclaimer

-0.68

LOC

-0.68

MODE

-0.67

POSITIVE LOGITS

 millennium

0.71

ope

0.69

omers

0.68

ight

0.68

areth

0.68

obi

0.67

anga

0.67

etus

0.67

 blasp

0.65

 loft

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.