INDEX

Explanations

affixes and morphological variations of words suggesting measurement or degrees

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

endings like -ment, -mente

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

 ainfi

-0.93

 auroit

-0.90

 chrétien

-0.89

 anún

-0.89

 étoient

-0.89

 avoient

-0.88

 pouvoit

-0.86

 feroit

-0.83

 plufieurs

-0.80

berdayakan

-0.79

POSITIVE LOGITS

jmniej

0.57

0.56

 regent

0.56

(!)

0.55

Leary

0.53

(!)

0.53

ínű

0.52

FCS

0.52

 versa

0.51

 Nelly

0.51

Activations Density 0.016%

affixes and morphological variations of words suggesting measurement or degrees

endings like -ment, -mente

No Comments

No Known Activations