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

 monet

-0.81

 confid

-0.64

edient

-0.63

 condem

-0.62

ilater

-0.61

nec

-0.61

 spar

-0.61

 pedd

-0.61

 investing

-0.61

 plagiar

-0.60

POSITIVE LOGITS

aan

0.77

batch

0.74

 Dress

0.70

ushima

0.69

Ã©n

0.68

borg

0.67

pine

0.65

Ã§

0.64

illard

0.63

atonin

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.