INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

 arts

-0.70

 pinch

-0.68

ary

-0.67

 premium

-0.66

 sixteen

-0.64

uxe

-0.63

dra

-0.63

 coun

-0.63

 sized

-0.63

 projection

-0.62

POSITIVE LOGITS

\"

0.74

egg

0.71

agher

0.70

lege

0.69

ibble

0.68

roe

0.68

 Cantor

0.68

FTWARE

0.68

ACP

0.68

 Rebell

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.