INDEX

Explanations

the word "anyone" with a high activation value

oai_token-act-pair · gpt-3.5-turbo

mention of the word "anyone."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 Nanto

-0.66

 Barg

-0.62

 Congo

-0.61

 Cast

-0.61

efficient

-0.60

ritz

-0.59

pa

-0.59

 Maze

-0.59

 Hound

-0.59

 Labor

-0.58

POSITIVE LOGITS

 else

1.51

THING

1.24

Else

1.10

omever

0.99

 Else

0.99

soever

0.97

else

0.95

ĪĴ

0.88

zzle

0.87

 imaginable

0.83

Activations Density 0.016%

the word "anyone" with a high activation value

mention of the word "anyone."

No Comments

No Known Activations