INDEX

Explanations

words related to controversies, debates, and opinions

oai_token-act-pair · gpt-3.5-turbo

punctuated phrases that indicate equality or contrast in ideas

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

quished

-0.83

»Ĵ

-0.78

(@

-0.69

enary

-0.68

GGGG

-0.66

Ĥİ

-0.64

©¶æ

-0.64

ukong

-0.63

)!

-0.63

forth

-0.63

POSITIVE LOGITS

 whereas

1.05

but

1.00

 preferring

0.96

but

0.90

rather

0.84

 rather

0.83

ie

0.83

 implying

0.82

 although

0.80

 ignoring

0.78

Activations Density 0.570%

words related to controversies, debates, and opinions

punctuated phrases that indicate equality or contrast in ideas

No Comments

No Known Activations