INDEX

Explanations

phrases indicating relationships and actions of entities within a context

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

start of a sentence or clause

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

 kasarigan

-0.55

⤹

-0.49

Personendaten

-0.47

 Roskov

-0.44

 uska

-0.44

Попис

-0.43

 betweenstory

-0.41

Rohy

-0.40

ftagPool

-0.40

stateProvider

-0.40

POSITIVE LOGITS

Kjelder

0.44

SWT

0.43

 verwijzen

0.43

 Houſe

0.42

 برانيه

0.42

 Theſe

0.41

 insuffisamment

0.41

enovo

0.39

 houſe

0.39

Tembelea

0.39

Activations Density 0.046%

phrases indicating relationships and actions of entities within a context

start of a sentence or clause

No Comments

No Known Activations