INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

undo

-0.76

lus

-0.73

 TODAY

-0.71

edom

-0.68

ursed

-0.67

 skeletons

-0.66

rew

-0.66

\/

-0.66

 Blessed

-0.65

Ëľ

-0.65

POSITIVE LOGITS

Elsewhere

0.82

 proble

0.79

Downloadha

0.79

Sort

0.71

Chart

0.69

Rated

0.68

Grade

0.63

efficients

0.62

cue

0.61

NPR

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.