INDEX

Explanations

interactions involving assistance and favors

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

helping others

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

setVerticalGroup

-0.53

Datuak

-0.43

="@+

-0.43

matchCondition

-0.40

endcsname

-0.40

MatIconModule

-0.39

ToolStrip

-0.38

AdapterView

-0.37

chränk

-0.37

 measured

-0.36

POSITIVE LOGITS

']>;

0.46

erapeutics

0.45

aiuto

0.44

########.

0.43

liferay

0.41

 Screws

0.41

:+:

0.41

 насељу

0.41

 pomocy

0.40

 membantu

0.40

Activations Density 0.030%

interactions involving assistance and favors

helping others

No Comments

No Known Activations