INDEX

Explanations

phrases indicating ordinal positions or rankings

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

ordinal and final/superlative

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

Portail

-0.51

ougou

-0.45

 laku

-0.44

 Reit

-0.41

DOC

-0.41

atguigu

-0.40

MathML

-0.40

ariat

-0.40

 deport

-0.40

soaked

-0.40

POSITIVE LOGITS

 second

0.51

 ultimate

0.48

 third

0.46

 extremely

0.45

 seemingly

0.45

 seventeenth

0.44

 arguably

0.44

 béisbol

0.44

 absolut

0.44

 favourite

0.44

Activations Density 0.025%

phrases indicating ordinal positions or rankings

ordinal and final/superlative

No Comments

No Known Activations