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

uve

-0.69

 Boone

-0.67

 Lionel

-0.65

 Jere

-0.65

 Juven

-0.65

ocaly

-0.65

 Arri

-0.65

Coy

-0.64

IER

-0.64

ertodd

-0.63

POSITIVE LOGITS

DragonMagazine

0.84

RFC

0.79

cdn

0.77

netflix

0.76

tc

0.74

seless

0.70

illing

0.69

reads

0.68

 spoil

0.67

ham

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.