INDEX

Explanations

phrases related to procedural rules and consequences in formal contexts

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

will be consequences

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

 esternos

-0.59

setVerticalGroup

-0.56

fjspx

-0.53

tangentMode

-0.52

HideFlags

-0.49

lgari

-0.47

 المر

-0.46

 AttributeSet

-0.46

 Walkover

-0.46

ValueStyle

-0.46

POSITIVE LOGITS

EndContext

0.38



0.37

 considérons

0.32

 OGSÅ

0.31

战斗力

0.30

 revoked

0.30

 consider

0.30

তথ্যসূত্র

0.30

DockStyle

0.29

 will

0.29

Activations Density 0.016%

phrases related to procedural rules and consequences in formal contexts

will be consequences

No Comments

No Known Activations