INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

Jr

-0.79

ãĤ´ãĥ³

-0.77

byter

-0.71

kefeller

-0.67

 authorised

-0.66

gc

-0.65

 Slot

-0.65

 Instruct

-0.63

 Minister

-0.62

 advoc

-0.61

POSITIVE LOGITS

tery

0.76

ribune

0.74

icles

0.70

axies

0.67

pora

0.66

trop

0.65

ients

0.65

rations

0.64

ording

0.63

agne

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.