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

leck

-0.66

 Marketable

-0.65

âķĲâķĲ

-0.64

 Â¥

-0.60

ixture

-0.59

 wrapper

-0.59

bos

-0.59

illac

-0.59

ership

-0.58

 à¨

-0.58

POSITIVE LOGITS

annis

0.74

 Collider

0.73

sighted

0.73

minded

0.71

hid

0.68

intendent

0.68

atible

0.67

 photographers

0.67

attled

0.67

terday

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.