INDEX

Explanations

repetitive or common terms related to categories or identifiers in various contexts

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

unisex bathroom, plugins, purchase, Nexus tablet, Troubles

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

IntoConstraints

-0.62

PreInfinity

-0.59

GraphicsUnit

-0.55

 nød

-0.54

➟

-0.54

 wuß

-0.53

TagMode

-0.53

WriteTagHelper

-0.52

ьаж

-0.52

 متعلقه

-0.52

POSITIVE LOGITS

gle

0.41

Den

0.40

tar

0.39

••••

0.37

 minuti

0.36

Tar

0.35

 rehe

0.35

Pranala

0.33

Ang

0.33

 glance

0.33

Activations Density 0.005%

repetitive or common terms related to categories or identifiers in various contexts

unisex bathroom, plugins, purchase, Nexus tablet, Troubles

No Comments

No Known Activations