INDEX

Explanations

mentions of a specific term or phrase "Hum" appearing in various contexts

oai_token-act-pair · gpt-3.5-turbo

references to humor

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

"+

-0.70

 Peaks

-0.69

cape

-0.68

×Ķ

-0.63

ECTION

-0.62

 approve

-0.60

Adds

-0.60

é¾įå

-0.60

 Trader

-0.60

hips

-0.60

POSITIVE LOGITS

Hum

1.18

hum

1.12

ankind

0.97

orously

0.96

pty

0.92

ility

0.92

ouri

0.91

iday

0.89

iliation

0.89

Hum

0.88

Activations Density 0.008%

mentions of a specific term or phrase "Hum" appearing in various contexts

references to humor

No Comments

No Known Activations