INDEX

Explanations

references to prominent items or entities within a context

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

current state or options

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

toMatchSnapshot

-0.49

 pageContext

-0.41

حوالہ

-0.39

fromnode

-0.39

 GMAT

-0.38

HSE

-0.38

 jspb

-0.37

 Resume

-0.37

APM

-0.37

 Fiske

-0.36

POSITIVE LOGITS

 Infórmanos

0.62

 kasarigan

0.58

The

0.46

tigkeits

0.46

 chính

0.44

LookAnd

0.44

ihnachts

0.43

这款

0.42

 accompanying

0.42

The

0.42

Activations Density 0.007%

references to prominent items or entities within a context

current state or options

No Comments

No Known Activations