INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

TL

-0.78

 Plate

-0.71

 Derby

-0.69

 Presence

-0.69

 Bore

-0.68

 Chaser

-0.66

ker

-0.65

FER

-0.63

REAM

-0.63

RAG

-0.62

POSITIVE LOGITS

oÄŁ

0.78

ghai

0.74

tarian

0.71

 condem

0.71

hack

0.70

iveness

0.69

wip

0.68

Downloadha

0.66

 sidel

0.66

ilan

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.