INDEX

Explanations

complex mathematical notation and structured expressions

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

mathematical arrays and matrices

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

+#+

-0.80

GEBURTSDATUM

-0.59

 sumpay

-0.58

 INTERESAR

-0.56

 pinulongan

-0.54

➯

-0.53

 podjela

-0.52

twimg

-0.51

#+#

-0.49

Gweler

-0.48

POSITIVE LOGITS

uxxxx

0.46

 refirió

0.40

weightedMode

0.39

 continúas

0.39

 Buche

0.36

halb

0.36

 comenzará

0.35

InjectAttribute

0.34

 Weiß

0.34

FDRE

0.34

Activations Density 0.164%

complex mathematical notation and structured expressions

mathematical arrays and matrices

No Comments

No Known Activations