INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

å°Ĩ

-0.64

wn

-0.63

RF

-0.63

sung

-0.62

 Heroes

-0.62

oured

-0.60

uming

-0.59

vo

-0.59

ifter

-0.59

KNOWN

-0.59

POSITIVE LOGITS

ertodd

0.71

 Kobe

0.70

veyard

0.69

ument

0.68

 Kush

0.67

stad

0.67

eus

0.67

 Lama

0.67

 Haku

0.63

ļéĨĴ

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.