INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 Kard

-0.75

Aus

-0.68

BBC

-0.65

 Dare

-0.65

 Palest

-0.63

Austral

-0.63

payer

-0.61

 Yose

-0.59

Bal

-0.59

ajor

-0.59

POSITIVE LOGITS

rol

0.78

ogie

0.75

ãĤ´ãĥ³

0.68

ridge

0.66

keye

0.63

à©

0.61

ngth

0.61

rup

0.61

ogged

0.59

tin

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.