INDEX

Explanations

proper nouns or names, particularly those related to individuals

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

Japanese particles and gemini

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 31-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_31/width_131k/average_l0_109

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.31.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

prv

-0.49

 abor

-0.45

anf

-0.45

 Souver

-0.45

wijk

-0.44

 Augustus

-0.44

 cán

-0.44

 cari

-0.43

 Tulip

-0.43

spring

-0.42

POSITIVE LOGITS

े

0.58

 ThemeData

0.52

 gambe

0.52

 Waray

0.52

gemaakt

0.50

eb

0.48

 estatales

0.48

 fédé

0.47

 femininas

0.46

Activations Density 0.011%

proper nouns or names, particularly those related to individuals

Japanese particles and gemini

No Comments

No Known Activations