INDEX

Explanations

phrases related to processes or events reaching a conclusion or peak

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

culmination

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

 Minder

-0.49

 neuve

-0.45

 ско

-0.43

 nég

-0.42

 Económica

-0.41

 médicale

-0.41

 armée

-0.40

 Tasche

-0.40

tráfico

-0.40

 katze

-0.40

POSITIVE LOGITS

 culmination

1.05

 culminating

0.97

 culminated

0.92

 culmin

0.92

 climax

0.84

 crowning

0.71

 pinnacle

0.63

invokeLater

0.63

bootstrapcdn

0.59

 consummation

0.59

Activations Density 0.009%

phrases related to processes or events reaching a conclusion or peak

culmination

No Comments

No Known Activations