INDEX

Explanations

specific numeric values, descriptions or identifiers within general contexts

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

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 snippetHide

-0.84

httphttps

-0.72

XmlAccessorType

-0.67

хьтан

-0.63

\{\\

-0.62

UserScript

-0.60

፩

-0.58

 createSprite

-0.58

exitRule

-0.57

 surla

-0.56

POSITIVE LOGITS

 Constitucional

0.35

Zalety

0.34

ölcs

0.33

Better

0.33

 Herzens

0.32

 better

0.32

 нравится

0.31

 rotating

0.31

 constitucional

0.31

 מוס

0.31

Activations Density 0.246%

specific numeric values, descriptions or identifiers within general contexts

No Comments

No Known Activations