INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

oster

-0.66

 earnest

-0.64

ãĤ¹

-0.62

itious

-0.62

Neo

-0.61

 ..............

-0.60

 lowly

-0.59

knit

-0.58

allel

-0.58

assic

-0.58

POSITIVE LOGITS

BOOK

0.76

raf

0.73

 confir

0.68

zees

0.66

eat

0.64

trial

0.64

 Democr

0.61

ocr

0.61

Marginal

0.61

oÄŁ

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.