INDEX

Explanations

phrases related to trends or directions

oai_token-act-pair · gpt-3.5-turbo

expressions related to trends and following movements

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

ayne

-0.66

onge

-0.65

arer

-0.63

wed

-0.61

ongyang

-0.61

 Regist

-0.59

 refere

-0.59

 banned

-0.59

onal

-0.58

aband

-0.58

POSITIVE LOGITS

leon

0.79

DragonMagazine

0.75

 financially

0.72

cale

0.70

chnology

0.69

 creatively

0.69

 bandwagon

0.67

 Dickinson

0.67

 Carney

0.66

":[

0.65

Activations Density 0.210%

phrases related to trends or directions

expressions related to trends and following movements

No Comments

No Known Activations