INDEX

Explanations

quotation marks and statements within quotes

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

quoted phrases and their explanations

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

})*/

-0.39

 vôtre

-0.38

rungsseite

-0.38

 feroit

-0.37

üğ

-0.36

 Numerade

-0.36

 vêtement

-0.35

 propOrder

-0.35

dialement

-0.34

 tjen

-0.34

POSITIVE LOGITS

Portale

0.63

 kaynağından

0.57

⎣

0.56

 nawr

0.55

ourdes

0.55

Datuak

0.55

}{*}{

0.54

`.`

0.53

chiede

0.51

 كومونز

0.51

Activations Density 0.026%

quotation marks and statements within quotes

quoted phrases and their explanations

No Comments

No Known Activations