INDEX

Explanations

keywords and phrases related to research and information gathering

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

questions about information

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

 Italijanski

-0.61

 Administrativna

-0.60

 oprot

-0.57

 Roskov

-0.57

arakhand

-0.55

agascar

-0.53

 bezeichneter

-0.52

SharedDtor

-0.52

 Gotham

-0.49

writeFieldEnd

-0.49

POSITIVE LOGITS

 Infórmanos

0.47

Lähteet

0.47

ällor

0.42

 récente

0.42

Fazit

0.40

 auprès

0.39

 mayoría

0.38

tidas

0.38

 Khusus

0.38

 Głów

0.37

Activations Density 0.040%

keywords and phrases related to research and information gathering

questions about information

No Comments

No Known Activations