INDEX

Explanations

phrases indicating locations or geographical descriptions

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

location names after of or in

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

aarrggbb

-0.52

-0.48

 raiſ

-0.47

 leaſt

-0.46

 Anſ

-0.45

 Inſ

-0.44

upyter

-0.41

WebElementEntity

-0.41

 anſ

-0.41

ſtance

-0.40

POSITIVE LOGITS

Portail

0.56

 ExecuteAsync

0.48

homonymie

0.48

ⓧ

0.47

 southeastern

0.46

 ThemeData

0.45

 southeast

0.44

GEBURTS

0.43

Décès

0.43

 autorytatywna

0.43

Activations Density 0.024%

phrases indicating locations or geographical descriptions

location names after of or in

No Comments

No Known Activations