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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

orse

-0.80

ãģ®éŃĶ

-0.80

}}}

-0.79

lest

-0.65

{"

-0.63

itatively

-0.62

 Sorce

-0.61

 therefore

-0.61

itative

-0.61

WHY

-0.60

POSITIVE LOGITS

ado

0.78

obin

0.72

aukee

0.67

atown

0.66

gage

0.66

ikuman

0.64

 impunity

0.63

EMENT

0.63

 likeness

0.63

Khe

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.