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

enger

-0.73

XL

-0.72

ng

-0.71

SEA

-0.66

nell

-0.65

Nap

-0.65

nect

-0.65

kies

-0.64

iT

-0.64

nel

-0.63

POSITIVE LOGITS

bush

0.83

 captcha

0.79

thinkable

0.69

 gobl

0.68

bris

0.67

pmwiki

0.66

bent

0.65

escription

0.64

 damned

0.63

Reviewer

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.