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

 mathemat

-0.93

cius

-0.83

ciating

-0.73

ãĤ·ãĥ£

-0.71

ombs

-0.67

rador

-0.67

quickShipAvailable

-0.67

ummies

-0.67

utra

-0.67

ctica

-0.67

POSITIVE LOGITS

 insur

0.72

Stre

0.68

 Sabha

0.64

 Grimes

0.61

Ye

0.60

itary

0.60

Dir

0.58

ĥ

0.58

Strength

0.57

Minor

0.56

Activations Density 0.000%

No Known Activations

This feature has no known activations.