INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

acebook

-0.68

 skirm

-0.66

efeated

-0.65

quartered

-0.61

 Tall

-0.61

 subsequent

-0.60

 Archdemon

-0.60

ammad

-0.60

 semblance

-0.60

 Pearce

-0.60

POSITIVE LOGITS

çīĪ

0.78

)]

0.75

Catholic

0.72

è¦

0.72

Downloadha

0.69

Pope

0.68

]"

0.68

ãĤĮ

0.67

displayText

0.66

)</

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.