INDEX

Explanations

phrases related to movie trailers or promotional material

oai_token-act-pair · gpt-3.5-turbo

references to "The Force Awakens" and related teasers

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

mel

-0.73

inges

-0.72

Murray

-0.68

Ü

-0.65

Laf

-0.63

parts

-0.62

UM

-0.62

 Wass

-0.62

ogical

-0.61

scribed

-0.60

POSITIVE LOGITS

 Awakens

1.47

akens

0.78

 Skywalker

0.75

qi

0.74

ignt

0.74

ignty

0.73

 Kenobi

0.70

smanship

0.70

terness

0.70

uggest

0.69

Activations Density 0.011%

phrases related to movie trailers or promotional material

references to "The Force Awakens" and related teasers

No Comments

No Known Activations