INDEX

Explanations

occurrences of mathematical terms and symbols related to set notation and sequences

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

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 utafitiHapana

-0.76

 للاسماء

-0.73

 Anſ

-0.66

 onely

-0.65

 avoient

-0.63

 betweenstory

-0.62

 nôtre

-0.60

astify

-0.60

 greateſt

-0.60

 محفوظة

-0.60

POSITIVE LOGITS

in

0.70

∈

0.66

∈

0.66

IN

0.57

In

0.57

In

0.56

in

0.56

notin

0.53

 within

0.48

isIn

0.48

Activations Density 0.078%

occurrences of mathematical terms and symbols related to set notation and sequences

No Comments

No Known Activations