INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

adr

-0.70

 entirety

-0.66

 debtor

-0.65

berus

-0.63

Chi

-0.63

theless

-0.63

abal

-0.63

 contra

-0.63

 Earthqu

-0.62

ysics

-0.61

POSITIVE LOGITS

enhagen

0.79

baum

0.71

Registered

0.68

heimer

0.68

enegger

0.66

berg

0.63

Stack

0.62

anski

0.60

 Seasons

0.60

 broadcasters

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.