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

 brill

-0.75

axter

-0.72

eatures

-0.72

ufact

-0.68

ERA

-0.67

gio

-0.66

redd

-0.64

ModLoader

-0.64

 mathemat

-0.63

gyn

-0.62

POSITIVE LOGITS

 slack

0.66

Sed

0.64

 there

0.63

 è£ıè¦ļéĨĴ

0.63

 Powell

0.60

TRE

0.60

ãĢı

0.60

 XCOM

0.58

 Qing

0.58

Bie

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.