INDEX

Explanations

patterns of numerical values or symbols that indicate lists or sequences

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 23-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_23/width_16k/average_l0_120

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.23.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

IUrlHelper

-0.55

issauga

-0.55

antası

-0.54

 Theſe

-0.53

 Schülerinnen

-0.52

 Bewußt

-0.52

 Majefty

-0.50

⟬

-0.50

 Chwiliwch

-0.50

 wikipagina

-0.49

POSITIVE LOGITS

 betweenstory

0.40

0.39

complexContent

0.39

 tramonto

0.35

 Light

0.35

 saluted

0.34

Ke

0.34

confidence

0.33

شهاد

0.33

 confidence

0.32

Activations Density 0.118%

patterns of numerical values or symbols that indicate lists or sequences

No Comments

No Known Activations