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

ËĪ

-0.89

itol

-0.87

Sov

-0.86

OPA

-0.72

auder

-0.70

Sov

-0.65

idon

-0.64

ivia

-0.64

ensis

-0.63

|--

-0.63

POSITIVE LOGITS

 collaborator

0.70

agnetic

0.68

mail

0.65

 theirs

0.64

awed

0.62

 dagger

0.60

sab

0.60

arsh

0.59

senal

0.59

 vitri

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.