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

VB

-0.71

tle

-0.70

abad

-0.69

Self

-0.69

byte

-0.66

 Savior

-0.64

Dem

-0.63

 è£ıç

-0.63

 Emin

-0.62

mission

-0.62

POSITIVE LOGITS

noon

0.71

 Saras

0.66

 knots

0.66

 Cthulhu

0.64

etheless

0.64

Ãį

0.63

 foreground

0.63

assic

0.63

rul

0.63

 crafts

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.