INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

headers

-0.93

mbuds

-0.77

esy

-0.72

mares

-0.71

ĵĺ

-0.70

 contrace

-0.68

ctors

-0.68

 impunity

-0.67

ife

-0.66

giene

-0.66

POSITIVE LOGITS

 Yuan

0.71

avorite

0.71

uno

0.69

odd

0.68

amel

0.67

>>\

0.66

Percent

0.64

ilial

0.64

GMT

0.63

NM

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.