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

VD

-0.84

eye

-0.78

ocamp

-0.77

meric

-0.75

andre

-0.73

von

-0.72

arie

-0.72

ussion

-0.71

iku

-0.69

gall

-0.68

POSITIVE LOGITS

 Arrows

0.77

 Commun

0.71

grave

0.69

 Spectre

0.66

 Jericho

0.66

 goodbye

0.65

 Harmony

0.65

 Goodbye

0.64

haps

0.63

 Bullets

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.