INDEX

Explanations

words or phrases associated with being in an opposite position or direction

oai_token-act-pair · gpt-3.5-turbo

instances of the word "opposite" and related phrases indicating opposing directions or contrasts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

ule

-0.83

ebook

-0.81

urrent

-0.81

haw

-0.78

ourning

-0.78

ulic

-0.77

zinski

-0.77

ortium

-0.76

beans

-0.75

ourn

-0.74

POSITIVE LOGITS

 sides

1.01

 directions

0.91

 sexes

0.86

osite

0.83

 side

0.82

 oppos

0.81

 direction

0.79

 halves

0.78

 twins

0.78

 twin

0.78

Activations Density 0.018%

words or phrases associated with being in an opposite position or direction

instances of the word "opposite" and related phrases indicating opposing directions or contrasts

No Comments

No Known Activations