INDEX

Explanations

the word "linda" with varying activations

oai_token-act-pair · gpt-3.5-turbo

occurrences of the name "Linda."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 frost

-0.71

 platforms

-0.64

 draft

-0.60

 safely

-0.60

 vault

-0.59

 rated

-0.57

 Snap

-0.57

post

-0.57

 Champions

-0.57

/+

-0.57

POSITIVE LOGITS

inda

4.35

issa

1.22

antha

1.17

indu

1.15

inia

1.15

ind

1.13

uti

1.08

ydia

1.06

INTON

1.06

 Linda

1.05

Activations Density 0.014%

the word "linda" with varying activations

occurrences of the name "Linda."

No Comments

No Known Activations