INDEX

Explanations

actions or behaviors that involve doing something

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

"do" followed by a pronoun or noun

np_acts-logits-general · gemini-2.0-flash

do the something

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 22-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_22/width_16k/average_l0_72

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.22.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

tvguidetime

-0.63

Tembelea

-0.60

httphttps

-0.59

原始内容存档于

-0.57

Искәрмәләр

-0.57

########.

-0.56

 ProtoMessage

-0.55

ausea

-0.54

providedIn

-0.54

Derbyniad

-0.53

POSITIVE LOGITS

 everything

0.81

 whatever

0.79

 something

0.71

 exactly

0.70

 things

0.69

 what

0.68

 some

0.66

 away

0.57

 anything

0.57

 unspeakable

0.57

Activations Density 0.106%

actions or behaviors that involve doing something

"do" followed by a pronoun or noun

do the something

No Comments

No Known Activations