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

————

-0.97

minist

-0.72

rower

-0.71

achine

-0.70

Offline

-0.67

ells

-0.67

ound

-0.66

thur

-0.65

 Column

-0.63

Pool

-0.62

POSITIVE LOGITS

perture

0.85

inia

0.70

isa

0.69

nces

0.66

annis

0.66

 cake

0.66

ometers

0.65

phant

0.63

cence

0.62

ometry

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.