INDEX

Explanations

numerical data or codes

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

`A_{` with modifications

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

 CONGRESS

-0.56

 Tsch

-0.54

 DIFFERENT

-0.54

MessageTagHelper

-0.53

 Schuster

-0.52

 unheard

-0.52

ynthesis

-0.51

 wikihow

-0.51

ohistochemistry

-0.50

 THIRD

-0.50

POSITIVE LOGITS

 autorytatywna

0.51

 للاسماء

0.49

})*/

0.42

 beginnetje

0.40

})->

0.31

RTLI

0.30

 ujednoznacz

0.30

Kaynakça

0.30

})),

0.29

]},

0.28

Activations Density 0.000%

numerical data or codes

`A_{` with modifications

No Comments

No Known Activations