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

stadt

-0.76

rans

-0.65

NetMessage

-0.60

 interrupt

-0.60

nes

-0.60

Rez

-0.60

nor

-0.59

 Koen

-0.59

opian

-0.59

Member

-0.58

POSITIVE LOGITS

 Fenrir

0.72

hao

0.71

ruary

0.68

fts

0.67

owship

0.67

ibaba

0.66

 Kappa

0.65

 teasp

0.65

":"/

0.63

uese

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.