INDEX

Explanations

references to political positions and affiliations

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

official positions and audits

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

 protoimpl

-0.60

fjspx

-0.60

脚注の使い方

-0.53

DoubleQuotes

-0.50

 inesperado

-0.50

 CreateTagHelper

-0.48

UnusedPrivate

-0.47

DockStyle

-0.47

yntaxException

-0.47

AndEndTag

-0.47

POSITIVE LOGITS

 students

0.40

 jute

0.40

 pupils

0.39

timewa

0.39

 audits

0.38

Tut

0.38

 Moral

0.38

 utafitiHapana

0.38

 kasarigan

0.37

 Purg

0.37

Activations Density 0.022%

references to political positions and affiliations

official positions and audits

No Comments

No Known Activations