INDEX

Explanations

the word "know"

oai_token-act-pair · gpt-3.5-turbo

the phrase "you know" as a discourse marker

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

utenberg

-0.85

anmar

-0.77

stad

-0.75

omal

-0.71

iets

-0.71

upe

-0.70

otom

-0.69

cius

-0.68

ermanent

-0.66

aez

-0.65

POSITIVE LOGITS

lege

0.81

ledged

0.80

ledge

0.74

terday

0.73

 exactly

0.66

abella

0.66

how

0.65

 KNOW

0.64

 guessed

0.64

 whats

0.63

Activations Density 0.033%

the word "know"

the phrase "you know" as a discourse marker

No Comments

No Known Activations