INDEX

Explanations

constructs related to mathematical formatting or expressions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 5-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_5/width_16k/average_l0_68

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.5.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

 crock

-0.68

 Capp

-0.64

Kud

-0.64

 Madd

-0.63

 stell

-0.61

 Popp

-0.59

DEAN

-0.58

chila

-0.57

 Ramen

-0.57

 firework

-0.57

POSITIVE LOGITS

évaluateur

0.75

 prisonniers

0.73

 animato

0.72

 âgé

0.72

 financières

0.71

 cœurs

0.70

 olacaktır

0.69

 intervento

0.69

 enfans

0.68

 pulito

0.67

Activations Density 0.094%

constructs related to mathematical formatting or expressions

No Comments

No Known Activations