INDEX

Explanations

instances of the phrase "to" followed by an action or intention

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

come to understand

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

 tigre

-0.51

 myſelf

-0.51

 Templo

-0.51

 lópez

-0.50

 houſe

-0.50

 sánchez

-0.49

 Svensk

-0.49

 rodríguez

-0.48

 используя

-0.47

 fubject

-0.47

POSITIVE LOGITS

 getItemId

0.60

twimg

0.60

 come

0.60

come

0.55

CreateModel

0.54

 Administrativna

0.52

makeText

0.51

 Come

0.51

ContextHolder

0.50

 COME

0.49

Activations Density 0.010%

instances of the phrase "to" followed by an action or intention

come to understand

No Comments

No Known Activations