INDEX

Explanations

references to officers in various contexts

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

officer or officers

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

Big

-0.48

The

-0.44

 Haywood

-0.44

Ham

-0.44

big

-0.44

Big

-0.43

the

-0.42

 fechar

-0.41

 Wain

-0.40

 nowych

-0.40

POSITIVE LOGITS

 Officers

1.21

Officer

1.19

 officers

1.18

 officer

1.17

 Officer

1.17

officer

1.16

Officers

1.13

officers

1.12

 OFFICER

1.08

 OFFICERS

1.06

Activations Density 0.006%

references to officers in various contexts

officer or officers

No Comments

No Known Activations