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

CLS

-0.74

Pwr

-0.70

 Forth

-0.66

 farming

-0.65

sol

-0.64

 migration

-0.64

MAP

-0.63

 Accounting

-0.63

 revival

-0.63

 fortunes

-0.62

POSITIVE LOGITS

fight

0.84

iens

0.81

oug

0.79

Ã®

0.76

verbal

0.76

ividually

0.76

iott

0.75

cats

0.74

ittens

0.74

words

0.74

Activations Density 0.000%

No Known Activations

This feature has no known activations.