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

uden

-0.76

efully

-0.72

ropolis

-0.72

geon

-0.69

ried

-0.68

yles

-0.67

 gard

-0.67

 canv

-0.66

emies

-0.65

ource

-0.65

POSITIVE LOGITS

ĪĴ

0.80

Dat

0.78

 Shuttle

0.75

 Satellite

0.73

 Stain

0.69

 Identity

0.68

 Tablet

0.67

liter

0.67

 Instr

0.65

 Slot

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.