INDEX

Explanations

conjunctions and linking phrases that connect ideas or clauses

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

preceding articles or conjunctions

np_acts-logits-general · gemini-2.0-flash

conjunction with subsequent description

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 18-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_18/width_16k/average_l0_74

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.18.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

Aholisi

-0.57

 Arkivert

-0.56

 StringTokenizer

-0.55

 esternos

-0.54

actéristi

-0.51

Билгалдахарш

-0.51

versies

-0.51

 sentía

-0.51

 andato

-0.51

TypedDataSet

-0.50

POSITIVE LOGITS

 allows

1.23

 enables

1.22

 helps

1.20

 enable

1.11

 allow

1.05

 help

1.00

 prevents

0.99

 ensures

0.98

 reduces

0.98

helps

0.97

Activations Density 0.422%

conjunctions and linking phrases that connect ideas or clauses

preceding articles or conjunctions

conjunction with subsequent description

No Comments

No Known Activations