INDEX

Explanations

proper nouns related to locations or organizations

oai_token-act-pair · gpt-3.5-turbo

references to governmental or important institutional structures

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

hari

-0.83

yne

-0.83

iness

-0.80

gel

-0.76

iest

-0.75

ecycle

-0.66

ily

-0.66

uth

-0.65

ener

-0.65

ocative

-0.65

POSITIVE LOGITS

itol

0.84

OTUS

0.82

 sidx

0.79

 Mellon

0.75

 Kislyak

0.72

ippi

0.70

bledon

0.69

 Paramount

0.69

cit

0.68

 sten

0.68

Activations Density 0.022%

proper nouns related to locations or organizations

references to governmental or important institutional structures

No Comments

No Known Activations