INDEX

Explanations

phrases related to capability and performance

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

ability and how

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

IntoConstraints

-0.68

 disambiguazione

-0.62

ArgsConstructor

-0.62

verwijspagina

-0.57

AndEndTag

-0.57

 endblock

-0.56

Попис

-0.56

abetes

-0.52

WidgetItem

-0.51

moveToFirst

-0.50

POSITIVE LOGITS

どれだけ

0.48

 ability

0.47

 whether

0.41

PhysRev

0.40

how

0.40

 насколько

0.40

 effectiveness

0.38

いかに

0.35

MarshalTo

0.35

有多少

0.34

Activations Density 0.042%

phrases related to capability and performance

ability and how

No Comments

No Known Activations