INDEX

Explanations

the word "noticed."

oai_token-act-pair · gpt-3.5-turbo

instances of the word “notice” and its variations, indicating observations or awareness

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

quer

-0.73

prep

-0.73

 negotiator

-0.69

export

-0.67

cop

-0.67

wives

-0.66

cise

-0.66

ccording

-0.65

ãĥ³ãĤ¸

-0.65

venge

-0.65

POSITIVE LOGITS

how

0.78

cules

0.76

ury

0.73

 noticed

0.71

adow

0.68

lessly

0.66

 notices

0.64

flies

0.64

enance

0.63

 spikes

0.62

Activations Density 0.023%

the word "noticed."

instances of the word “notice” and its variations, indicating observations or awareness

No Comments

No Known Activations