INDEX

Explanations

references to procedural guidelines or instructions

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

sentences following "configures"

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

 obé

-0.39

 feroit

-0.35

 épar

-0.35

 curé

-0.34

 auroit

-0.34

 akku

-0.34

 chrétien

-0.34

 pouvoit

-0.33

 consommer

-0.32

 Gallimard

-0.31

POSITIVE LOGITS

NameInMap

0.73

transQ

0.72

 intptr

0.69

 propOrder

0.68

 autorytatywna

0.63

endregion

0.62

 disponibilités

0.62

.*")]

0.61

 ProtoMessage

0.60

蚪

0.60

Activations Density 0.124%

references to procedural guidelines or instructions

sentences following "configures"

No Comments

No Known Activations