INDEX

Explanations

phrases related to control settings or mechanisms

oai_token-act-pair · gpt-3.5-turbo

mentions of "controls," specifically related to gaming or systems management

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

 Apostle

-0.77

ographies

-0.72

ONY

-0.67

 apostle

-0.64

ony

-0.63

olog

-0.63

Rated

-0.63

Voy

-0.62

 Incarn

-0.62

ographic

-0.61

POSITIVE LOGITS

 controls

0.82

ulators

0.78

creen

0.76

rador

0.73

eers

0.72

 levers

0.70

 guarding

0.69

ELL

0.69

edin

0.69

eering

0.68

Activations Density 0.014%

phrases related to control settings or mechanisms

mentions of "controls," specifically related to gaming or systems management

No Comments

No Known Activations