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

olan

-0.68

ware

-0.66

 Robb

-0.64

ize

-0.64

strate

-0.63

angelo

-0.63

 draped

-0.62

ows

-0.61

ying

-0.60

anza

-0.59

POSITIVE LOGITS

dayName

0.85

etheless

0.69

responsible

0.67

raph

0.65

cephal

0.65

ccording

0.65

 Pastebin

0.64

IRE

0.64

 Hiroshima

0.62

Leon

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.