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

doi

-0.90

-0.82

Reviewer

-0.81

 Editors

-0.71

ussen

-0.69

ally

-0.69

SCP

-0.68

æ©Ł

-0.66

onym

-0.66

FactoryReloaded

-0.65

POSITIVE LOGITS

 Trident

0.67

marine

0.67

iator

0.62

osph

0.60

udding

0.60

mite

0.59

 cradle

0.59

qt

0.59

cture

0.58

chell

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.