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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

tle

-0.79

emort

-0.67

orers

-0.66

ession

-0.65

alty

-0.65

 chops

-0.65

escape

-0.65

edom

-0.64

fer

-0.64

ains

-0.63

POSITIVE LOGITS

Magikarp

0.73

Vari

0.71

 Ambro

0.64

Scientists

0.63

 Scientists

0.63

 Wink

0.62

andro

0.62

there

0.61

 Cogn

0.60

auld

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.