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

 Commons

-0.71

rophe

-0.69

IENT

-0.68

 Jong

-0.63

 Geological

-0.63

 tending

-0.62

PLIED

-0.62

 Bronx

-0.61

 Comics

-0.61

 Bangkok

-0.61

POSITIVE LOGITS

uyomi

0.97

ahime

0.85

:,

0.72

netflix

0.70

wagen

0.69

imize

0.69

buff

0.67

udeb

0.66

 hump

0.66

erest

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.