INDEX

Explanations

sequences or patterns in structured data representations

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

mathematical symbols like + and $

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

machung

-0.35

 گیری

-0.35

 Absch

-0.35

 Flucht

-0.35

-0.34

 surla

-0.34

de

-0.34

接

-0.34

uoja

-0.33

POSITIVE LOGITS

 незавершена

0.69

ConstraintMaker

0.69

 Vikipedi

0.67

parsedMessage

0.57

:+:

0.57

@"/

0.56

 AssemblyCulture

0.55

Портали

0.55

تقاوى

0.54

 typelib

0.54

Activations Density 0.000%

sequences or patterns in structured data representations

mathematical symbols like + and $

No Comments

No Known Activations