INDEX

Explanations

time-related expressions, specifically related to 24-hour periods

oai_token-act-pair · gpt-3.5-turbo

references to 24-hour time or continuous service

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

oslav

-0.75

ãĤ©

-0.69

 Eliot

-0.66

agogue

-0.64

etsk

-0.64

olini

-0.64

poor

-0.63

 poems

-0.62

ieties

-0.61

bom

-0.61

POSITIVE LOGITS

hour

0.98

0.88

hours

0.79

0.72

Hours

0.71

Hour

0.71

ournal

0.68

eware

0.66

Month

0.64

assian

0.64

Activations Density 0.094%

time-related expressions, specifically related to 24-hour periods

references to 24-hour time or continuous service

No Comments

No Known Activations