INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

Enlarge

-0.85

ADRA

-0.73

Mub

-0.72

Muk

-0.70

 MÃ¼

-0.67

PLIC

-0.66

 GOODMAN

-0.65

cffffcc

-0.65

Marginal

-0.65

 externalToEVAOnly

-0.65

POSITIVE LOGITS

ewitness

0.88

ories

0.80

die

0.73

ologically

0.71

ensable

0.71

ophobic

0.71

pole

0.71

 sheds

0.71

igible

0.70

bish

0.70

Activations Density 0.000%

No Known Activations

This feature has no known activations.