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

arium

-0.71

 Leaving

-0.68

 Runner

-0.65

 Patron

-0.64

 Minotaur

-0.64

BIP

-0.63

 Pizza

-0.61

abal

-0.60

 Torah

-0.59

 Neph

-0.57

POSITIVE LOGITS

ĪĴ

0.91

uilt

0.77

thora

0.75

lihood

0.73

unte

0.73

emis

0.72

 stereotype

0.70

onymous

0.69

andem

0.68

roup

0.68

Activations Density 0.000%

No Known Activations

This feature has no known activations.