INDEX

Explanations

phrases where someone is speaking to or directing words at another person

oai_token-act-pair · gpt-3.5-turbo

instances of the word "to" 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 @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

cyclop

-0.73

published

-0.72

umerable

-0.69

icut

-0.68

FORMATION

-0.68

motion

-0.67

ItemThumbnailImage

-0.66

swick

-0.65

 authored

-0.64

Includes

-0.64

POSITIVE LOGITS

 passers

1.03

 someone

0.96

 somebody

0.95

him

0.93

 superiors

0.89

 strangers

0.87

 subordinates

0.86

 whoever

0.83

me

0.80

 oneself

0.80

Activations Density 0.237%

phrases where someone is speaking to or directing words at another person

instances of the word "to" in various contexts

No Comments

No Known Activations