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

 submar

-0.70

prototype

-0.70

ĸļ

-0.70

 elector

-0.69

esses

-0.66

berra

-0.66

erg

-0.66

 mosquit

-0.65

ikuman

-0.65

istar

-0.64

POSITIVE LOGITS

GG

0.74

0.70

 evenly

0.67

(.

0.65

PLA

0.65

pan

0.65

bear

0.64

":[{"

0.64

score

0.63

Po

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.