INDEX

Explanations

emotive expressions and conversational discourse markers

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

Follows interjections ("well", "how", etc.)

np_acts-logits-general · gemini-2.0-flash

well followed by a word

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 24-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_24/width_16k/average_l0_73

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.24.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Knows

-0.63

]=>

-0.62

 Wikimedijinoj

-0.62



-0.60

'}),

-0.60

 Dont

-0.59

',(

-0.55

 pyplot

-0.55

iecie

-0.54

Dont

-0.54

POSITIVE LOGITS

don

1.03

let

0.99

 wait

0.99

why

0.97

 thank

0.94

 tell

0.94

 listen

0.92

how

0.91

 sorry

0.87

 remember

0.86

Activations Density 0.238%

emotive expressions and conversational discourse markers

Follows interjections ("well", "how", etc.)

well followed by a word

No Comments

No Known Activations