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

©¶æ

-0.77

 Templ

-0.69

 Nguyen

-0.68

Rs

-0.68

externalActionCode

-0.67

 Hunters

-0.66

ä½ľ

-0.63

 Huang

-0.63

Ì¶

-0.62

Cash

-0.61

POSITIVE LOGITS

rier

0.76

itage

0.75

awan

0.72

itia

0.72

abilia

0.71

ried

0.69

pse

0.69

phabet

0.68

}:

0.68

pas

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.