INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

 Nasa

-0.76

 Mankind

-0.68

SI

-0.67

XM

-0.67

 Breed

-0.67

 POLIT

-0.67

john

-0.66

HM

-0.65

 Mulcair

-0.65

Psych

-0.64

POSITIVE LOGITS

eport

0.78

lining

0.71

anchester

0.70

illin

0.69

flix

0.68

ership

0.67

uku

0.66

elist

0.65

 reorgan

0.65

vana

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.