INDEX

Explanations

instances of something being expected or intended to happen

oai_token-act-pair · gpt-3.5-turbo

phrases indicating expectations or obligations

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

lake

-0.77

lite

-0.74

cos

-0.70

cest

-0.70

tex

-0.69

croft

-0.69

cloth

-0.69

cards

-0.67

pu

-0.67

bu

-0.67

POSITIVE LOGITS

DonaldTrump

0.85

 explan

0.76

 culprit

0.68

 disclaim

0.68

STDOUT

0.67

 adjud

0.66

 conflic

0.66

 disappear

0.65

 unintention

0.63

asma

0.63

Activations Density 0.012%

instances of something being expected or intended to happen

phrases indicating expectations or obligations

No Comments

No Known Activations