INDEX

Explanations

references to administrative actions and legal terms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 40-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_40/width_16k/average_l0_125

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.40.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

chauung

-0.41

enseits

-0.41

DockStyle

-0.41

 vanuit

-0.38

prepareStatement

-0.37

 Directly

-0.37

thren

-0.36

ślę

-0.36

 igenom

-0.36

OGND

-0.36

POSITIVE LOGITS

 altogether

1.85

 entirely

1.54

 completely

1.40

 kokona

1.08

completely

1.08

 úplně

1.06

 Completely

1.05

 Altogether

1.05

 forever

1.04

 totally

1.03

Activations Density 0.908%

references to administrative actions and legal terms

No Comments

No Known Activations