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

schild

-0.92

steen

-0.88

ossier

-0.79

zee

-0.76

utterstock

-0.72

 cryptoc

-0.70

 Cathy

-0.68

aez

-0.67

udence

-0.67

Ã«

-0.66

POSITIVE LOGITS

HL

0.75

CHA

0.65

 Iron

0.62

ICAL

0.59

hest

0.59

 monog

0.58

ravel

0.57

Thumbnail

0.57

ali

0.56

OIL

0.56

Activations Density 0.000%

No Known Activations

This feature has no known activations.