INDEX

Explanations

details related to forum navigation and content organization

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

forum links and web pages

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

 surla

-0.41

연

-0.33

India

-0.33

the

-0.33

risti

-0.31

 routinely

-0.31

 counter

-0.31

 precisely

-0.30

 India

-0.30

Counter

-0.30

POSITIVE LOGITS

 AssemblyTitle

0.84

ngdoc

0.76

adaptiveStyles

0.73

 defaultstate

0.73

Geplaatst

0.72

enderror

0.68

 utafitiHapana

0.68

BASEPATH

0.67

Rujuakan

0.66

 AssemblyCompany

0.64

Activations Density 0.028%

details related to forum navigation and content organization

forum links and web pages

No Comments

No Known Activations