INDEX

Explanations

terms related to measurement and comparison

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

initial, beginning, start

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

 gyhoeddwyd

-0.66

mybatisplus

-0.58

complexContent

-0.54

endphp

-0.49

readFileSync

-0.48

 AssemblyCompany

-0.47

CodeAttribute

-0.46

GEBURTSDATUM

-0.46

 autorytatywna

-0.46

 Infórmanos

-0.45

POSITIVE LOGITS

 initial

0.83

 initially

0.81

initial

0.77

 Initial

0.73

 anfangs

0.71

 Initially

0.69

 iniziale

0.69

Initially

0.68

Initial

0.67

 INITIAL

0.65

Activations Density 0.113%

terms related to measurement and comparison

initial, beginning, start

No Comments

No Known Activations