INDEX

Explanations

frequent use of demonstrative pronouns and references to shared concepts

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

specific context or following words

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

ValueStyle

-0.66

uxxxx

-0.59

 ब्रेकडाउन

-0.56

 AssemblyCulture

-0.56

<unused47>

-0.55

<unused51>

-0.54

 resourceCulture

-0.54

<unused41>

-0.54

<unused14>

-0.54

<unused28>

-0.54

POSITIVE LOGITS

ContentLoaded

0.36

 hefyd

0.32

 protesta

0.32

 què

0.30

 összefoglaló

0.29

 emphasized

0.29

men

0.29

ArrowToggle

0.29

WebServlet

0.28

主要是

0.28

Activations Density 0.001%

frequent use of demonstrative pronouns and references to shared concepts

specific context or following words

No Comments

No Known Activations