INDEX

Explanations

words related to specific locations, typically universities and institutions

oai_token-act-pair · gpt-3.5-turbo

prepositions, particularly "in," used in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 incumb

-0.71

 innoc

-0.68

raf

-0.67

drivers

-0.65

 gaping

-0.65

DragonMagazine

-0.65

 defect

-0.63

ĺħ

-0.63

 unemploy

-0.62

POSITIVE LOGITS

 conjunction

1.34

 lieu

1.16

 accordance

1.03

 partnership

0.96

 relation

0.96

 addition

0.95

 Pasadena

0.92

 southeastern

0.91

ked

0.91

 Arlington

0.90

Activations Density 0.221%

words related to specific locations, typically universities and institutions

prepositions, particularly "in," used in various contexts

No Comments

No Known Activations