INDEX

Explanations

phrases that highlight irony or absurdity

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

before "too" expressing excess

np_acts-logits-general · gemini-2.0-flash

phrases describing poor writing

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

critical evaluations of media that call out contrivance or unrealistic, overly neat/predictable elements, often marked by intensifiers and evaluative qualifiers.

oai_token-act-pair · gpt-5 Triggered by @vetterc0

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.20.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

fjspx

-0.49

 Réponses

-0.48

 الحره

-0.44

 myſelf

-0.43

 ſta

-0.42

 leaſt

-0.41

EventId

-0.41

 faſt

-0.40

drawal

-0.40

 ſtate

-0.40

POSITIVE LOGITS

 güneş

0.40

XMLSchema

0.38

HideFlags

0.37

 zelve

0.37

uxxxx

0.37

 useRouter

0.37

OGND

0.37

 Constructors

0.36

msgTypes

0.36

 مرئيه

0.36

Activations Density 0.284%

phrases that highlight irony or absurdity

before "too" expressing excess

phrases describing poor writing

critical evaluations of media that call out contrivance or unrealistic, overly neat/predictable elements, often marked by intensifiers and evaluative qualifiers.

No Comments

No Known Activations