INDEX

Explanations

occurrences of the word "Mad" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 2-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_2/width_16k/average_l0_67

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.2.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

 présidenti

-0.35

WithIOException

-0.34

 urbain

-0.33

maakt

-0.33

 énergé

-0.33

Доброго

-0.32

 ambas

-0.32

 metodología

-0.32

 chercheurs

-0.32

 ambos

-0.31

POSITIVE LOGITS

 justify

0.81

 justified

0.80

 justifies

0.77

ContentLoaded

0.75

mad

0.74

Mad

0.74

justified

0.73

Mad

0.71

Jus

0.70

Jus

0.69

Activations Density 0.060%

occurrences of the word "Mad" in various contexts

No Comments

No Known Activations