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

qqa

-0.73

rament

-0.70

hov

-0.70

bucks

-0.68

ollen

-0.65

Reviewer

-0.64

arrow

-0.63

hari

-0.63

hog

-0.62

eu

-0.60

POSITIVE LOGITS

format

0.66

DEV

0.64

 unpop

0.62

 regress

0.61

append

0.61

uits

0.60

iscons

0.59

ength

0.59

Jackson

0.59

igure

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.