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

lihood

-0.78

adequ

-0.73

 ital

-0.72

 instincts

-0.68

çİĭ

-0.67

 calculus

-0.66

©¶æ

-0.65

 situational

-0.64

 defences

-0.63

 defenses

-0.63

POSITIVE LOGITS

natureconservancy

0.81

ickr

0.80

Gaza

0.79

UGE

0.77

iddler

0.75

facebook

0.70

urred

0.69

enced

0.68

UG

0.68

ogi

0.68

Activations Density 0.000%

No Known Activations

This feature has no known activations.