INDEX

Explanations

definite and indefinite articles indicating nouns

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

become a

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

 Treue

-0.44

 nobre

-0.41

-0.38

 Verhält

-0.38

 Noyes

-0.37

 Dokumentation

-0.35

 Volkes

-0.35

",

-0.35

 Milán

-0.34

POSITIVE LOGITS

become

1.05

 become

1.01

 Become

0.95

becomes

0.92

 becomes

0.91

Become

0.90

became

0.85

 Becomes

0.84

 became

0.84

 Became

0.84

Activations Density 0.011%

definite and indefinite articles indicating nouns

become a

No Comments

No Known Activations