INDEX

Explanations

specific identifiers and naming conventions within technical or programming contexts

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

initials and abbreviations

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

 ―――――

-0.74

 queſta

-0.73

NameInMap

-0.65

Ꞓ

-0.64

 queſto

-0.63

$_(

-0.63

 CanadaChoose

-0.62

<unused0>

-0.61

 Jefus

-0.60

 ſind

-0.60

POSITIVE LOGITS

 exportación

0.31

bewerken

0.31

 importación

0.28

 venganza

0.28

 manio

0.28

 exportaciones

0.28

 dignidad

0.27

 conmigo

0.27

 aussi

0.26

 också

0.26

Activations Density 0.081%

specific identifiers and naming conventions within technical or programming contexts

initials and abbreviations

No Comments

No Known Activations