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

agall

-0.69

Parts

-0.67

Destroy

-0.62

Materials

-0.60

zz

-0.60

 Divinity

-0.59

cffffcc

-0.58

 Punch

-0.57

odder

-0.57

Scroll

-0.57

POSITIVE LOGITS

inois

0.77

 Trout

0.67

IPA

0.66

 ALEC

0.64

lihood

0.64

Sym

0.62

yip

0.62

IPM

0.61

 Appalachian

0.61

imity

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.