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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Į

-0.72

Ĵ

-0.72

 Cheong

-0.71

roo

-0.71

ı

-0.69

ãĥ¼ãĥĨ

-0.64

 Heal

-0.64

 Singer

-0.64

Ļ

-0.64

 "$:/

-0.63

POSITIVE LOGITS

 residue

0.71

iasm

0.70

achus

0.68

irements

0.68

ization

0.67

 digit

0.66

asm

0.65

ebted

0.64

 depended

0.63

oresc

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.