INDEX

Explanations

phrases that denote approval, denial, and compliance in the context of decision-making

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

actions or states

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

 GenerationType

-0.49

contentLoaded

-0.47

erapeutics

-0.46

 ilha

-0.44

AlterField

-0.44

合い

-0.42

sula

-0.42

itale

-0.42

 Hutchins

-0.41

četně

-0.41

POSITIVE LOGITS

 ویکی‌پدیای

0.46

AccessorTable

0.43

complexContent

0.42

 estekak

0.42

Autoritní

0.39

 betweenstory

0.39

BuildContext

0.38

+#+#

0.38

 مشين

0.36

 otomatig

0.36

Activations Density 0.153%

phrases that denote approval, denial, and compliance in the context of decision-making

actions or states

No Comments

No Known Activations