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

inally

-0.71

estern

-0.69

ast

-0.67

Äį

-0.66

asting

-0.65

aste

-0.65

ardon

-0.64

gs

-0.63

enta

-0.62

"+

-0.62

POSITIVE LOGITS

BuyableInstoreAndOnline

0.77

erville

0.74

 answ

0.72

ELF

0.70

 bund

0.69

arsh

0.68

 newsp

0.67

é»Ĵ

0.67

 unden

0.67

 juven

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.