INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

UNCH

-0.68

ãĥ¢

-0.65

 Chronicle

-0.64

 Collider

-0.63

 funn

-0.61

IVES

-0.60

EED

-0.60

EXT

-0.60

 bananas

-0.60

 Zika

-0.60

POSITIVE LOGITS

tel

0.81

Ħ¢

0.77

iances

0.71

existing

0.68

?]

0.67

isted

0.66

bands

0.66

galitarian

0.66

aum

0.66

held

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.