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

 bells

-0.83

 lamps

-0.70

 headlights

-0.69

 toys

-0.67

Mom

-0.66

 trim

-0.65

 Toys

-0.64

 figur

-0.64

 blinking

-0.63

 cartoon

-0.63

POSITIVE LOGITS

SG

0.75

eln

0.74

eneg

0.74

/(

0.71

traumatic

0.70

*/(

0.70

irtual

0.69

eva

0.69

illus

0.68

+(

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.