INDEX

Explanations

a variety of specific nouns related to biology, education, and health contexts

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

medical terms and geographical locations

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

MENAFN

-0.52

hab

-0.42

Rohy

-0.42

vacy

-0.41

bac

-0.40

المناصب

-0.39

兼

-0.39

TypedDataSet

-0.38

/**

-0.38

}],

-0.38

POSITIVE LOGITS

vœ

0.54

 quelcon

0.51

 Económica

0.49

 étoient

0.46

 LIRE

0.45

 desejo

0.45

 typelib

0.45

 boeken

0.44

ambién

0.44

 dessus

0.44

Activations Density 0.036%

a variety of specific nouns related to biology, education, and health contexts

medical terms and geographical locations

No Comments

No Known Activations