INDEX

Explanations

words indicating certainty or predictions about outcomes

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

also including/containing

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

 vectorstock

-0.47

EndGlobalSection

-0.46

цезда

-0.46

GOTREF

-0.44

 вікі

-0.44

adaptiveStyles

-0.43

ínű

-0.43

 alem

-0.42

 AssemblyCompany

-0.42

ymal

-0.42

POSITIVE LOGITS

 includes

0.74

 include

0.69

 شامل

0.66

 contains

0.66

 mencakup

0.61

 contain

0.59

 bevat

0.57

 encompass

0.57

 also

0.56

 itself

0.56

Activations Density 0.052%

words indicating certainty or predictions about outcomes

also including/containing

No Comments

No Known Activations