INDEX

Explanations

phrases indicating actions, states, or conditions related to influence and progression

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

Follows certain adverbs/conjunctions

np_acts-logits-general · gemini-2.0-flash

also because, also include

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-16k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

Sunlight

-0.51

fin

-0.49

carbon

-0.49

udsman

-0.48

-0.47

 Execution

-0.47

fen

-0.47

allocate

-0.47

Execution

-0.47

mol

-0.47

POSITIVE LOGITS

 '\\;'

0.41

()")

0.37

 tampoco

0.37

 CURIAM

0.36

 duquel

0.36

 propOrder

0.34

PreferredItem

0.34

 επίσης

0.34

zelfde

0.34

?“

0.33

Activations Density 2.231%

phrases indicating actions, states, or conditions related to influence and progression

Follows certain adverbs/conjunctions

also because, also include

No Comments

No Known Activations