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

oÄŁ

-0.82

ouf

-0.82

omsky

-0.80

aughters

-0.76

ooks

-0.76

ocalypse

-0.76

trop

-0.73

akov

-0.73

lees

-0.72

ourt

-0.70

POSITIVE LOGITS

 Brand

0.65

 âĸº

0.65

enger

0.64

 Theft

0.64

 seizure

0.62

ience

0.62

âĢ¢âĢ¢

0.62

Sne

0.60

 badge

0.60

 gesture

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.