INDEX

Explanations

phrases that involve the use of the word "dubbed" or "labeled" in various contexts

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

naming or labeling

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

 églises

-0.46

 prêtres

-0.43

 policiers

-0.40

 peuple

-0.39

 inget

-0.39

 colère

-0.39

 Comune

-0.39

 Juifs

-0.39

 plancher

-0.39

 actualizaciones

-0.38

POSITIVE LOGITS

 dubbed

0.81

 AssemblyTitle

0.75

 moniker

0.71

 transfieras

0.69

 labeled

0.69

 coined

0.68

 labelled

0.67

 branded

0.67

makeText

0.65

 termed

0.65

Activations Density 0.019%

phrases that involve the use of the word "dubbed" or "labeled" in various contexts

naming or labeling

No Comments

No Known Activations