INDEX

Explanations

frequent and critical data points and references in a technical context

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

data stream equivalent

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

fortawesome

-0.54

msgTypes

-0.49

 Савезне

-0.45

 crossorigin

-0.45

InjectAttribute

-0.44

 Baillargeon

-0.44

 ويكيميديا

-0.44

pim

-0.44

PFC

-0.43

 jScrollPane

-0.43

POSITIVE LOGITS

 mitään

0.53

 jugado

0.51

 plegable

0.50

 ciudadana

0.49

 suaminya

0.48

nehå

0.46

 kysy

0.46

TagMode

0.46

 Moscú

0.46

 Económica

0.46

Activations Density 0.032%

frequent and critical data points and references in a technical context

data stream equivalent

No Comments

No Known Activations

frequent and critical data points and references in a technical context

data stream equivalent

No Comments

No Known Activations