INDEX

Explanations

references to existential states or outcomes

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

code blocks and symbols

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

rungsseite

-0.61

 poveznice

-0.59

 مشين

-0.50

ArrowToggle

-0.48

paravant

-0.45

Autoritní

-0.45

 Италијани

-0.45

 GenerationType

-0.44

 capucha

-0.43

perator

-0.42

POSITIVE LOGITS

出版年

0.42

}();

0.39

 InputDecoration

0.37

 ")");

0.37

(!__

0.36

Pid

0.36

 onCreateView

0.35

RTLU

0.35

findpost

0.35

psy

0.35

Activations Density 0.105%

references to existential states or outcomes

code blocks and symbols

No Comments

No Known Activations