INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

################

-0.81

 unfocusedRange

-0.81

Cow

-0.79

KR

-0.74

TIT

-0.70

 dstg

-0.69

EXT

-0.69

KK

-0.65

artifacts

-0.64

GG

-0.63

POSITIVE LOGITS

icum

0.80

aceous

0.74

enne

0.71

ngth

0.70

isi

0.70

 Dialogue

0.69

naires

0.68

tained

0.67

acea

0.67

 agre

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.