INDEX

Explanations

references to quoted text and navigational phrases

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

blog post metadata

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

angliski

-0.47

 gover

-0.46

queryInterface

-0.45

ględ

-0.39

 individual

-0.39

LLocation

-0.39

Personendaten

-0.39

変

-0.38

 others

-0.37

кипедия

-0.37

POSITIVE LOGITS

 Numerade

0.73

 nahilalakip

0.60

 AssemblyProduct

0.58

mappedBy

0.57

 "..\..\

0.53

XmlAccessType

0.50

VersionUID

0.49

 "..\..\..\

0.48

labelledby

0.47

0.46

Activations Density 0.004%

references to quoted text and navigational phrases

blog post metadata

No Comments

No Known Activations