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

CRIP

-0.77

ertodd

-0.76

Maps

-0.76

 Pastebin

-0.72

utenberg

-0.71

 Perse

-0.70

ENSE

-0.69

Frag

-0.68

á½

-0.68

affles

-0.67

POSITIVE LOGITS

board

1.12

boards

0.89

isle

0.69

used

0.67

ahime

0.65

 Ezek

0.65

 Targ

0.65

hai

0.65

boarding

0.64

 sheltered

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.