INDEX

Explanations

instances of emphasis or strong statements

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

answer and responde

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

featureID

-0.72

 transfieras

-0.52

 auroit

-0.50

EndGlobalSection

-0.50

DoubleQuotes

-0.50

Jeografia

-0.48

 dalamnya

-0.48

 يتيمه

-0.47

msgTypes

-0.47

fören

-0.46

POSITIVE LOGITS

 answer

0.48

answer

0.42

<code>

0.42

回答

0.41

答

0.40

 utafitiHapana

0.40

<blockquote>

0.40

Answer

0.39

 responde

0.38

 answers

0.38

Activations Density 0.154%

instances of emphasis or strong statements

answer and responde

No Comments

No Known Activations