INDEX

Explanations

phrases related to numerical quantities and distinct entities

oai_token-act-pair · gpt-3.5-turbo

references to quantifiable entities or categories

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

brance

-0.81

heit

-0.78

matter

-0.72

govtrack

-0.71

culosis

-0.69

etermination

-0.67

abad

-0.66

-0.64

otto

-0.64

life

-0.64

POSITIVE LOGITS

 aforementioned

0.95

 pillars

0.91

 platforms

0.83

ses

0.81

 avenues

0.81

 eras

0.81

 bedrooms

0.80

 earliest

0.79

 extremes

0.79

 occasions

0.79

Activations Density 0.126%

phrases related to numerical quantities and distinct entities

references to quantifiable entities or categories

No Comments

No Known Activations