INDEX

Explanations

instances of the word "might" indicating possibility

oai_token-act-pair · gpt-3.5-turbo

expressions of possibility or conjecture

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

coord

-0.59

arget

-0.59

eger

-0.57

 Period

-0.56

abet

-0.56

geries

-0.55

oup

-0.55

UTC

-0.55

heat

-0.54

raham

-0.54

POSITIVE LOGITS

 might

3.02

might

2.62

 Might

2.03

may

2.02

 could

1.56

 ought

1.52

 would

1.42

could

1.39

MAY

1.34

 wouldn

1.28

Activations Density 0.040%

instances of the word "might" indicating possibility

expressions of possibility or conjecture

No Comments

No Known Activations