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

hyde

-0.84

ngth

-0.80

DragonMagazine

-0.79

maxwell

-0.77

Ĭ±

-0.74

pha

-0.74

POSE

-0.74

ģĸ

-0.74

¬¼

-0.74

POSITIVE LOGITS

 managed

0.82

rified

0.63

 theoretical

0.63

 awaited

0.63

0.62

tl

0.61

 contro

0.60

 hinted

0.59

Twe

0.59

 confirmed

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.