INDEX

Explanations

words or phrases that indicate emphasis or importance

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

introduces a topic or action

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

MigrationBuilder

-0.68

ConstraintMaker

-0.59

LookAnd

-0.45

outWeight

-0.42

fromnode

-0.40

 optimized

-0.40

 defaultstate

-0.39

postsleuth

-0.39

LoadScene

-0.38

 betweenstory

-0.38

POSITIVE LOGITS

OGND

0.59

\{\\

0.56

ViewImports

0.52

oneofs

0.49

 EconPapers

0.47

("")]

0.47

 BIBLE

0.46

rungsseite

0.46

 desmotivaciones

0.45

RegistryLite

0.45

Activations Density 0.000%

words or phrases that indicate emphasis or importance

introduces a topic or action

No Comments

No Known Activations

words or phrases that indicate emphasis or importance

introduces a topic or action

No Comments

No Known Activations