INDEX

Explanations

programming-related syntax and structure, especially involving labels and categories

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_20/width_16k/average_l0_71

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

 FetchType

-0.51

oneofs

-0.43

gel

-0.41

mouseup

-0.41

 przeciw

-0.40

ষ্

-0.39

 http

-0.39

ঝ

-0.39

 bậc

-0.39

 kvä

-0.38

POSITIVE LOGITS

WebVitals

0.96

AndEndTag

0.89

Vidite

0.88

CloseOperation

0.86

 ſeveral

0.84

Controllo

0.83

 Theſe

0.82

oredCriteria

0.77

enumi

0.75

 theſe

0.75

Activations Density 0.074%

programming-related syntax and structure, especially involving labels and categories

No Comments

No Known Activations