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

eton

-0.91

ocide

-0.91

etics

-0.83

enary

-0.79

apons

-0.78

etically

-0.78

utra

-0.77

ukong

-0.75

enta

-0.75

owship

-0.73

POSITIVE LOGITS

 Clubs

0.69

 Aven

0.66

 Fare

0.66

 goodbye

0.65

 magn

0.63

bill

0.62

 demanding

0.62

 actionGroup

0.61

 Icar

0.60

 clubs

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.