INDEX

Explanations

comparative and superlative phrases indicating varying degrees of intensity or quantity

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

following "as", "too", or "more"

np_acts-logits-general · gemini-2.0-flash

as, more, too + extent/manner

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-16k

Configuration

google/gemma-scope-9b-it-res/layer_31/width_16k/average_l0_76

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

providedIn

-0.50

BeginContext

-0.48

částí

-0.39

homonymie

-0.37

 retours

-0.36

 ausreich

-0.36

 écoulé

-0.36

 größer

-0.36

Activité

-0.36

ftagPool

-0.35

POSITIVE LOGITS

 quickly

0.65

 forcefully

0.63

 extensively

0.63

 efficiently

0.63

 differently

0.62

 effectively

0.62

SourceChecksum

0.60

 Савезне

0.59

 prominently

0.56

 successfully

0.56

Activations Density 0.231%

comparative and superlative phrases indicating varying degrees of intensity or quantity

following "as", "too", or "more"

as, more, too + extent/manner

No Comments

No Known Activations