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

puff

-0.76

 Tasmania

-0.76

keye

-0.72

obos

-0.70

Ã¡n

-0.65

 Freeze

-0.64

oise

-0.63

hammer

-0.61

pal

-0.60

 Border

-0.60

POSITIVE LOGITS

emetery

0.91

itialized

0.89

 charism

0.71

 millenn

0.71

 cryptoc

0.70

ument

0.69

aughtered

0.67

ĸļ

0.67

llular

0.66

OHN

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.