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

kefeller

-0.73

WithNo

-0.69

natureconservancy

-0.68

guyen

-0.67

 welf

-0.67

 captcha

-0.64

thumbnails

-0.64

 tradem

-0.62

pired

-0.62

choes

-0.61

POSITIVE LOGITS

1.07

0.94

',

0.81

0.80

?,

0.77

ly

0.74

'),

0.70

,'

0.70

']

0.69

,[

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.