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

 Macedonia

-0.73

 Opera

-0.70

oglu

-0.67

FTA

-0.66

 Maced

-0.66

 Thumbnails

-0.66

IBLE

-0.64

XD

-0.63

 Hyundai

-0.62

otiation

-0.62

POSITIVE LOGITS

inch

0.74

."[

0.71

erville

0.71

rick

0.71

rix

0.69

shire

0.67

auri

0.67

arl

0.66

oise

0.66

rican

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.