INDEX

Explanations

words related to actions or states of being in various contexts, particularly those signifying relationships or conditions

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

restricted clauses following 'which'

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

 décé

-0.43

 Vordergrund

-0.41

bibinfo

-0.40

 polig

-0.39

Palmarès

-0.38

 Roskov

-0.38

 nhiêu

-0.38

jalá

-0.37

 kaynağından

-0.37

 nemlig

-0.36

POSITIVE LOGITS

 FetchType

0.55

 SwitchCompat

0.46

 Normdatei

0.44

+#+#

0.44

 protoimpl

0.43

IntoConstraints

0.41

AnimationsModule

0.41

jsii

0.40

IsContent

0.40

 capaces

0.39

Activations Density 0.090%

words related to actions or states of being in various contexts, particularly those signifying relationships or conditions

restricted clauses following 'which'

No Comments

No Known Activations