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

orrow

-0.76

atters

-0.70

olics

-0.70

 Mehran

-0.69

 Canberra

-0.68

 Compton

-0.67

Cath

-0.67

 Brisbane

-0.66

arella

-0.66

edIn

-0.65

POSITIVE LOGITS

 increment

0.76

position

0.70

ativity

0.68

backer

0.68

ftime

0.67

wagon

0.64

alph

0.64

 equity

0.63

gregation

0.63

llor

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.