INDEX

Explanations

variables, assignments, and directives related to programming structures

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

parentheses and code-like structures

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

 transfieras

-0.67

 AssemblyTitle

-0.65

 gynhyrchwyd

-0.63

 المعيارى

-0.62

thansa

-0.61

migrationBuilder

-0.60

TagMode

-0.57

tinyos

-0.57

 Chwiliwch

-0.56

webElementXpaths

-0.56

POSITIVE LOGITS

 Ausführung

0.34

 alternativ

0.34

 equivalent

0.34

 less

0.33

absolute

0.33

 just

0.32

Absolute

0.32

 atacado

0.32

 oculta

0.32

 alike

0.32

Activations Density 0.102%

variables, assignments, and directives related to programming structures

parentheses and code-like structures

No Comments

No Known Activations