INDEX

Explanations

code and programming-related terms, specifically regarding function declarations and database queries

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

technical nouns, citations, code

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

Blu

-0.34

Ro

-0.32

 <<<<<<<<<<<<<<

-0.31

-0.30

ุ

-0.29

sand

-0.28

Ben

-0.28

Al

-0.27

data

-0.27

Clo

-0.26

POSITIVE LOGITS

adaptiveStyles

0.71

 queſta

0.69

Datuak

0.68



0.68

SpringBootTest

0.63

 समीक्षक

0.62

FormTagHelper

0.60

 ویکی‌پدی

0.60

oneofs

0.59

 linkovi

0.59

Activations Density 0.256%

code and programming-related terms, specifically regarding function declarations and database queries

technical nouns, citations, code

No Comments

No Known Activations