INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 Mori

-0.73

URL

-0.73

im

-0.72

âĸĪâĸĪâĸĪâĸĪâĸĪâĸĪâĸĪâĸĪ

-0.69

awed

-0.67

 stumbling

-0.66

 ..........

-0.62

Jar

-0.62

NetMessage

-0.62

 dict

-0.61

POSITIVE LOGITS

enses

0.69

bye

0.68

 Pegasus

0.67

oso

0.66

obook

0.65

hiba

0.64

 Leilan

0.64

dule

0.63

thens

0.63

 absorbs

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.