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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

vals

-0.80

orah

-0.80

anguage

-0.78

meier

-0.77

heimer

-0.76

intel

-0.75

emort

-0.74

apego

-0.74

jri

-0.70

lain

-0.69

POSITIVE LOGITS

Tac

0.77

POP

0.71

 Quinn

0.71

 intest

0.64

 Tina

0.63

TED

0.62

 Flesh

0.62

cture

0.62

HIP

0.62

 Lima

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.