INDEX

Explanations

terms related to consequences and punitive actions

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

penalties and restrictions

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

AndEndTag

-0.65

 المعيارى

-0.49

 Spuren

-0.45

 desconocido

-0.44

 gyhoeddwyd

-0.43

 acuerdos

-0.42

 establecido

-0.41

/*

-0.40

MemoryWarning

-0.40

expandindo

-0.40

POSITIVE LOGITS

RegistryLite

0.54

новниш

0.44

ban

0.41

Демографія

0.40

future

0.40

 future

0.40

InputTagHelper

0.37

avenir

0.37

 access

0.36

 Oxide

0.36

Activations Density 0.043%

terms related to consequences and punitive actions

penalties and restrictions

No Comments

No Known Activations