INDEX

Explanations

references to various types of software components or systems

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

following "ark", "arm", "carb", or "erg"

np_acts-logits-general · gemini-2.0-flash

ark Matter

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-16k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

</table>

-0.61

Tó

-0.54

"}")

-0.54

]}\

-0.54

#>

-0.53

|<\

-0.52

}<\

-0.52

Soci

-0.51

 Obrador

-0.51

`<

-0.50

POSITIVE LOGITS

 ร์

0.60

rer

0.57

 borboleta

0.54

 Harrell

0.53

ror

0.51

 désert

0.50

 oreille

0.50

 vermelha

0.50

rr

0.50

 Larsson

0.50

Activations Density 1.543%

references to various types of software components or systems

following "ark", "arm", "carb", or "erg"

ark Matter

No Comments

No Known Activations