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

 ACTIONS

-0.84

Accessory

-0.70

Di

-0.68

 CLIENT

-0.68

Inv

-0.68

 Razer

-0.67

 Medieval

-0.66

uous

-0.65

Spanish

-0.63

translation

-0.63

POSITIVE LOGITS

arcity

0.85

ctors

0.82

 compress

0.73

 lengths

0.73

 airports

0.72

 Leban

0.71

 shelves

0.71

 reservoirs

0.69

awatts

0.67

 tremend

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.