INDEX

Explanations

phrases conveying surprise

oai_token-act-pair · gpt-3.5-turbo

expressions of surprise

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

•layer 5 - surprise

No Comments

Negative Logits

ciplinary

-0.84

amins

-0.74

vertisement

-0.73

haar

-0.73

utf

-0.71

ngth

-0.70

alach

-0.69

ueller

-0.69

interstitial

-0.69

ignty

-0.69

POSITIVE LOGITS

 Squid

0.79

imaru

0.74

Pew

0.72

vale

0.70

cules

0.70

 HuffPost

0.68

 bystanders

0.66

 onlook

0.66

 silence

0.65

Gry

0.65

Activations Density 0.023%

phrases conveying surprise

expressions of surprise

No Comments

No Known Activations