INDEX

Explanations

references to people's perceptions and social interactions

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

pronouns for people and groups

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

 ferien

-0.44

gher

-0.43

 besluit

-0.42

:✨

-0.41

Portion

-0.41



-0.41

 asti

-0.41

 nahilalakip

-0.40

ataan

-0.40

arote

-0.40

POSITIVE LOGITS

 audiences

0.48

 people

0.45

 strangers

0.45

 ProtoMessage

0.44

 audience

0.43

PhysRevLett

0.41

contentLoaded

0.41

 witnesses

0.40

fjspx

0.40

 others

0.39

Activations Density 0.054%

references to people's perceptions and social interactions

pronouns for people and groups

No Comments

No Known Activations