INDEX

Explanations

specific references to people, their roles, or attributes in a narrative context

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

r kelly agreed

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

markers of specific entities in expository noun phrases, especially pronouns/possessives and determiners tied to names or numeric details.

oai_token-act-pair · gpt-5 Triggered by @vetterc0

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

ientôt

-0.54

RegistryLite

-0.46

 ELECTRO

-0.46

 spolup

-0.45

Viitteet

-0.45

 humaine

-0.45

 Menschheit

-0.45

ismet

-0.45

 Вікіпе

-0.44

VERTISEMENT

-0.43

POSITIVE LOGITS

CWE

0.45

GLFW

0.42

StackNavigator

0.42

TabIndex

0.39

reportWebVitals

0.38

 quæ

0.37

 MonoBehaviour

0.36

もしれない

0.35

 Normdatei

0.34

')],

0.34

Activations Density 0.062%

specific references to people, their roles, or attributes in a narrative context

r kelly agreed

markers of specific entities in expository noun phrases, especially pronouns/possessives and determiners tied to names or numeric details.

No Comments

No Known Activations

specific references to people, their roles, or attributes in a narrative context

r kelly agreed

markers of specific entities in expository noun phrases, especially pronouns/possessives and determiners tied to names or numeric details.

No Comments

No Known Activations