INDEX

Explanations

phrases indicating assistance, improvement, and the potential for positive outcomes in various contexts

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

helping or implying outcomes

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

 surla

-0.57

sidemargin

-0.56

XmlRootElement

-0.53

StoryboardSegue

-0.52

IUrlHelper

-0.52

DockStyle

-0.51

 CURIAM

-0.49



-0.49

 resourceCulture

-0.48

featureID

-0.47

POSITIVE LOGITS

 further

0.52

 also

0.51

 dessutom

0.48

 additionally

0.46

更是

0.44

 außerdem

0.44

 moreover

0.43

さらに

0.42

 inoltre

0.42

 added

0.41

Activations Density 0.048%

phrases indicating assistance, improvement, and the potential for positive outcomes in various contexts

helping or implying outcomes

No Comments

No Known Activations