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

ãĥ³ãĤ¸

-0.68

plates

-0.65

etter

-0.65

esh

-0.63

 Emacs

-0.62

ãĤ¦

-0.61

idon

-0.60

avage

-0.59

 Morty

-0.59

lam

-0.58

POSITIVE LOGITS

CLASSIFIED

0.69

psey

0.68

 fuelled

0.67

 McDonnell

0.67

funding

0.67

 Seeking

0.66

=-=-=-=-=-=-=-=-

0.66

unity

0.64

IED

0.64

Experience

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.