INDEX

Explanations

phrases indicating lack of response or lack of confirmation

oai_token-act-pair · gpt-3.5-turbo

the occurrence of the word "did" 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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

case

-0.74

stood

-0.72

lined

-0.69

fields

-0.66

 Dise

-0.65

liner

-0.65

lining

-0.64

hog

-0.63

armed

-0.62

)=(

-0.62

POSITIVE LOGITS

actic

1.09

pez

0.99

 indeed

0.91

 manage

0.91

not

0.86

anos

0.83

ppel

0.82

 succeed

0.81

Finish

0.80

herty

0.79

Activations Density 0.095%

phrases indicating lack of response or lack of confirmation

the occurrence of the word "did" in various contexts

No Comments

No Known Activations