INDEX

Explanations

references to specific companies, publications, or notable locations

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

place names and brands

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

Jîn

-0.67

 betweenstory

-0.54

ThroughAttribute

-0.54

inchilla

-0.54

raborty

-0.53

 GenerationType

-0.52

دانشنامهٔ

-0.52

ennan

-0.52

urations

-0.50

cuterie

-0.50

POSITIVE LOGITS

 meurt

0.45

 Wikiseite

0.44

 évêque

0.43

 čierna

0.39

 ennemi

0.39

 CrossRef

0.38

 voluntarios

0.36

Soorten

0.36

 isolé

0.35

drawiam

0.35

Activations Density 0.255%

references to specific companies, publications, or notable locations

place names and brands

No Comments

No Known Activations