INDEX

Explanations

articles, specifically the word "a" and its variations

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

a/an followed by a word

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

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

 Administrativna

-0.66

лтемелер

-0.64

NameInMap

-0.64

FunctionFlags

-0.63

 lenker

-0.60

WindowConstants

-0.58

 Reſ

-0.56

 ſta

-0.56

 ProtoMessage

-0.55

Personendaten

-0.55

POSITIVE LOGITS

 currently

0.38

is

0.37

值为

0.36

is

0.36

 senior

0.36

 Klage

0.35

 deeply

0.34

gql

0.34

Activations Density 0.048%

articles, specifically the word "a" and its variations

a/an followed by a word

No Comments

No Known Activations

articles, specifically the word "a" and its variations

a/an followed by a word

No Comments

No Known Activations