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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

hra

-0.67

helm

-0.66

pta

-0.66

meg

-0.64

 incom

-0.63

 manifold

-0.62

ulf

-0.60

edo

-0.60

zig

-0.60

-0.58

POSITIVE LOGITS

quickShipAvailable

0.88

DERR

0.75

soDeliveryDate

0.74

READ

0.71

OPE

0.68

runners

0.67

ANG

0.64

DragonMagazine

0.63

ĨĴ

0.62

Virgin

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.