INDEX

Explanations

descriptions of unique products and their features

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

product brochures

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Configuration

google/gemma-scope-9b-pt-res/layer_24/width_131k/average_l0_97

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.24.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

 bordada

-0.31

 bordados

-0.31

 déjeuner

-0.30

 Acero

-0.30

 social

-0.29

 mulus

-0.29

 Trasp

-0.29

rätt

-0.29

 derecho

-0.28

 Derecho

-0.28

POSITIVE LOGITS

httphttps

0.72

⟬

0.72

tonsoft

0.71

Personendaten

0.70

хьтан

0.69

########.

0.69

writeFieldEnd

0.68

tagHelperRunner

0.66

MessageTagHelper

0.65

 Meksiku

0.65

Activations Density 0.020%

descriptions of unique products and their features

product brochures

No Comments

No Known Activations