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

earchers

-0.75

rontal

-0.72

oyal

-0.71

confidence

-0.71

obil

-0.70

ilitary

-0.69

activate

-0.66

quiet

-0.65

etimes

-0.65

rolet

-0.64

POSITIVE LOGITS

 cinem

0.68

seq

0.66

 americ

0.65

ost

0.64

 mont

0.64

iframe

0.62

 balloons

0.62

 Camer

0.62

 MILL

0.62

 balloon

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.