INDEX

Explanations

references to records or accounts

oai_token-act-pair · gpt-3.5-turbo

end-of-text markers and variations of the term "rec."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

MENTS

-0.77

ahime

-0.76

 flawed

-0.64

 flaws

-0.63

furt

-0.62

)=(

-0.59

falls

-0.59

holders

-0.58

MENT

-0.57

THING

-0.57

POSITIVE LOGITS

ruit

1.24

ycle

1.24

ipient

1.20

ords

1.18

urrent

1.18

ruits

1.17

ursive

1.16

ording

1.16

urring

1.09

secut

1.08

Activations Density 0.021%

references to records or accounts

end-of-text markers and variations of the term "rec."

No Comments

No Known Activations