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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

actionDate

-0.83

 separ

-0.78

 preference

-0.77

 identification

-0.69

 Czech

-0.67

sis

-0.67

 foreskin

-0.66

 respons

-0.66

ibl

-0.64

 compatibility

-0.64

POSITIVE LOGITS

ãĥīãĥ©

0.77

Reloaded

0.74

0.73

trap

0.69

iasco

0.68

Beast

0.68

POP

0.67

fleet

0.67

ultz

0.67

~~~~~~~~

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.