INDEX

Explanations

elements related to user information and data processing in a technical context

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

beginning of user input

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

.*")]

-0.37

 consistently

-0.35

 jalan

-0.31

 Ambiental

-0.30

ẹo

-0.30

➯

-0.29

 msglen

-0.29

能力

-0.29

 Polres

-0.29

 theatres

-0.29

POSITIVE LOGITS

Tembelea

0.78

 PhpStorm

0.77

Datuak

0.71

 Chwiliwch

0.70

httphttps

0.65

 للاسماء

0.61

 gynhyrchwyd

0.61

intellij

0.60

Geplaatst

0.59

ResumeLayout

0.57

Activations Density 0.000%

elements related to user information and data processing in a technical context

beginning of user input

No Comments

No Known Activations