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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Dalton

-0.77

omach

-0.76

hew

-0.71

 Anthem

-0.70

Ont

-0.66

ãĢ

-0.64

 seaf

-0.64

iii

-0.63

 Towns

-0.63

 Bachelor

-0.63

POSITIVE LOGITS

ulner

0.82

 guiActiveUn

0.78

paio

0.75

ascus

0.75

orsche

0.74

osal

0.74

icas

0.73

per

0.73

fman

0.73

wagen

0.72

Activations Density 0.000%

No Known Activations

This feature has no known activations.