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

pressed

-0.71

 Raider

-0.70

naires

-0.66

 Collider

-0.66

val

-0.65

 cour

-0.64

 comet

-0.63

 Cour

-0.62

Cay

-0.60

 Leilan

-0.59

POSITIVE LOGITS

agonists

0.70

illac

0.61

 Modes

0.61

iances

0.59

gag

0.59

iphate

0.58

okia

0.58

Girl

0.58

hene

0.58

æ©Ł

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.