INDEX

Explanations

comparative questions regarding numerical or value-based relationships

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

which is greater/smaller

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

 AssemblyCompany

-0.63

 wireType

-0.61

ckså

-0.59

########.

-0.57

 veľmi

-0.57

 very

-0.56

featureID

-0.56

 velmi

-0.55

majánló

-0.54

 hyvin

-0.54

POSITIVE LOGITS

 whichever

0.69

 between

0.59

どちら

0.58

 dominant

0.50

 Between

0.50

どっち

0.50

哪个

0.48

 BETWEEN

0.48

hichever

0.48

Between

0.47

Activations Density 0.103%

comparative questions regarding numerical or value-based relationships

which is greater/smaller

No Comments

No Known Activations