INDEX

Explanations

phrases related to announcements or introductions

oai_token-act-pair · gpt-3.5-turbo

references to collective identity or community ownership

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

puff

-0.76

conom

-0.75

cum

-0.72

bender

-0.72

tar

-0.66

more

-0.66

 belongs

-0.64

atican

-0.64

netflix

-0.64

 remembers

-0.63

POSITIVE LOGITS

selves

1.27

own

1.00

 respective

0.85

 collective

0.85

 newest

0.85

 asses

0.84

 motto

0.83

 dear

0.83

 beloved

0.81

 fearless

0.79

Activations Density 0.114%

phrases related to announcements or introductions

references to collective identity or community ownership

No Comments

No Known Activations