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

luaj

-0.78

;;;;;;;;;;;;

-0.77

ī

-0.74

BILITIES

-0.72

peria

-0.71

etheless

-0.67

ĨĴ

-0.66

CN

-0.65

 ðŁĺ

-0.65

loo

-0.64

POSITIVE LOGITS

CSI

0.72

 Saban

0.70

 disse

0.69

ablishment

0.68

 Tampa

0.66

apses

0.64

 mills

0.63

utenberg

0.63

 Hamp

0.62

 Tucker

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.