INDEX

Explanations

complex structured data patterns

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

structure, CUDA, SQL, legal, math, multilingual

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

ItemBackground

-0.82

iNdEx

-0.67

DIPSETTING

-0.66

 Infórmanos

-0.66

enumii

-0.65

Selfer

-0.62

 betweenstory

-0.62

kenstock

-0.57

 kasarigan

-0.56

GraphicsUnit

-0.56

POSITIVE LOGITS

 embarazadas

0.35

 wnętr

0.34

 Públicas

0.33

 parede

0.33

 ebenfalls

0.33

"){

0.32

 embarazada

0.32

)");

0.30

Ligações

0.29

樣子

0.29

Activations Density 0.015%

complex structured data patterns

structure, CUDA, SQL, legal, math, multilingual

No Comments

No Known Activations