INDEX

Explanations

instances of configuration-related terminology

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

configuration files

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

bootstrapcdn

-0.55

GEBURTSDATUM

-0.46

setVerticalGroup

-0.46

adaptiveStyles

-0.45

 سكانية

-0.45

+#+#

-0.44

الإنجليزية

-0.44

MethodManager

-0.43

AndroidJUnit

-0.43

fromnode

-0.42

POSITIVE LOGITS

conf

1.22

Conf

0.93

 conf

0.91

CONF

0.86

 Conf

0.84

 CONF

0.84

 Confessions

0.63

 configuration

0.57

 Configuration

0.57

 Confession

0.56

Activations Density 0.000%

instances of configuration-related terminology

configuration files

No Comments

No Known Activations