INDEX

Explanations

references related to the term "Wizard"

oai_token-act-pair · gpt-3.5-turbo

references to wizards and wizardry

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

PER

-0.76

uters

-0.76

rontal

-0.76

-0.72

andro

-0.69

pora

-0.68

DonaldTrump

-0.68

 sqor

-0.66

¿½

-0.66

tera

-0.65

POSITIVE LOGITS

ry

0.99

sonian

0.92

 wizard

0.89

ery

0.85

esses

0.83

loo

0.80

izard

0.77

fly

0.76

 Wizard

0.75

acular

0.73

Activations Density 0.020%

references related to the term "Wizard"

references to wizards and wizardry

No Comments

No Known Activations