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

©¶æ

-0.76

£ı

-0.70

ĪĴ

-0.70

 repeat

-0.69

yip

-0.68

ourke

-0.67

 mosqu

-0.67

ãĤ¨ãĥ«

-0.66

ryu

-0.66

 clipboard

-0.65

POSITIVE LOGITS

)</

0.74

lishes

0.73

iculty

0.70

avorable

0.68

haven

0.65

orse

0.64

escription

0.63

df

0.62

 Writers

0.60

Wyr

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.