INDEX

Explanations

phrases indicating design intent or target audience

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 21-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_21/width_16k/average_l0_129

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.21.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

 initComponents

-0.47

rungsseite

-0.45

MessageTagHelper

-0.44

 laikā

-0.43

 jadx

-0.41

 Audiodateien

-0.41

 båda

-0.39

RenderAtEndOf

-0.39

 terlibat

-0.37

 históricos

-0.37

POSITIVE LOGITS

 intended

0.83

 suited

0.79

Intended

0.77

 Intended

0.75

intended

0.72

 meant

0.71

 предназна

0.70

meant

0.69

 dành

0.68

suited

0.68

Activations Density 0.352%

phrases indicating design intent or target audience

No Comments

No Known Activations