INDEX

Explanations

references to people and their actions in various contexts

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

time expressions and code

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

Autoritní

-0.55

 AssemblyCulture

-0.54

ConstraintMaker

-0.52

SharedCtor

-0.51

wapV

-0.51

hematics

-0.50

protoimpl

-0.49

 للاسماء

-0.48

 Biôgrafia

-0.47

bourhood

-0.47

POSITIVE LOGITS

gridx

0.35

ticker

0.30

投

0.29

wx

0.29

classes

0.29

'.$

0.28

 Seele

0.28

目

0.28

GEBURTSDATUM

0.28

참고

0.28

Activations Density 0.129%

references to people and their actions in various contexts

time expressions and code

No Comments

No Known Activations