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

sic

-0.66

Tat

-0.66

isin

-0.65

overed

-0.64

Lad

-0.63

 Tibetan

-0.63

 Signed

-0.57

dom

-0.57

ãĥ¯

-0.57

 holders

-0.57

POSITIVE LOGITS

kinson

0.85

mington

0.75

heim

0.71

icter

0.71

 impart

0.70

©¶æ

0.68

APD

0.68

abase

0.67

 unconsciously

0.67

emia

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.