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

Tai

-0.71

 Constantin

-0.69

Tot

-0.67

alog

-0.67

Jud

-0.67

 Elliott

-0.66

obos

-0.65

 Thiel

-0.65

 advoc

-0.65

 Barrett

-0.65

POSITIVE LOGITS

azeera

0.80

onday

0.78

ppe

0.75

Enough

0.74

aneous

0.72

agate

0.71

heat

0.71

face

0.70

dar

0.67

rawdownloadcloneembedreportprint

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.