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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Zub

-0.75

selage

-0.69

wing

-0.62

zag

-0.61

 hinge

-0.61

 Chaff

-0.60

fit

-0.60

ological

-0.60

cov

-0.59

Wheel

-0.58

POSITIVE LOGITS

 Armory

0.74

RAM

0.73

OWN

0.69

agall

0.68

DI

0.66

ilage

0.66

ategory

0.64

usage

0.63

ovie

0.61

 Arcane

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.