INDEX

Explanations

patterns of zero or non-zero activations indicative of responses to mathematical factorization or divisibility questions

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

is x a multiple of y

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

 AssemblyCulture

-0.85

OGND

-0.74

 gynhyrchwyd

-0.71

MigrationBuilder

-0.70

 الرياضيه

-0.69

ConstraintMaker

-0.67

jspb

-0.65

 queſta

-0.63

 zwiſchen

-0.63

AddHtmlAttribute

-0.63

POSITIVE LOGITS

 again

0.41

 another

0.36

again

0.35

Another

0.34

 yine

0.33

Again

0.33

 Larsen

0.33

static

0.32

還有

0.32

Activations Density 0.037%

patterns of zero or non-zero activations indicative of responses to mathematical factorization or divisibility questions

is x a multiple of y

No Comments

No Known Activations