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

ãģ®éŃĶ

-0.75

 Inqu

-0.68

asus

-0.66

 UCHIJ

-0.66

profit

-0.66

swer

-0.65

 Scient

-0.64

ãĤ¡

-0.64

Greek

-0.63

 Political

-0.63

POSITIVE LOGITS

ortment

0.70

ijn

0.69

inators

0.68

 clipboard

0.66

 aspirations

0.65

allery

0.62

unders

0.62

alert

0.61

Parent

0.61

oaded

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.