INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

Ħ¢

-0.81

ftime

-0.78

roman

-0.71

ħĭ

-0.67

vl

-0.64

 Morales

-0.64

lda

-0.63

ngth

-0.62

inyl

-0.62

 Wast

-0.62

POSITIVE LOGITS

ees

0.68

ocate

0.65

ocations

0.64

 grasping

0.64

poral

0.62

avery

0.62

 atmosp

0.61

ogy

0.59

sense

0.59

 âĢķ

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.