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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 tremend

-0.70

fen

-0.69

 heap

-0.65

 whim

-0.64

 metab

-0.62

ilant

-0.60

 foreigner

-0.59

 shenan

-0.59

 born

-0.59

 councillors

-0.59

POSITIVE LOGITS

EMBER

0.80

JV

0.78

UV

0.77

QUIRE

0.77

WAYS

0.77

STEM

0.73

 srfAttach

0.72

uties

0.71

cd

0.71

VT

0.70

Activations Density 0.000%

No Known Activations

This feature has no known activations.