INDEX

Explanations

references to public information and its accessibility

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

publicly available information

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

 diría

-0.37

 consideró

-0.37

 partida

-0.36

 consideran

-0.35

 complementos

-0.34

 amores

-0.34

Tracce

-0.33

 spécialiste

-0.33

 conflictos

-0.33

 experto

-0.32

POSITIVE LOGITS

 publicly

0.93

 public

0.87

 Public

0.83

Public

0.83

 PUBLIC

0.80

 publick

0.76

 pública

0.75

PUBLIC

0.71

 öffentlich

0.70

 publice

0.69

Activations Density 0.034%

references to public information and its accessibility

publicly available information

No Comments

No Known Activations