INDEX

Explanations

specific tools or devices and their descriptions

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

describing shape and structure

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

httphttps

-0.46

 David

-0.44

David

-0.40

 initData

-0.38

 pimenta

-0.38

 productive

-0.38



-0.37

 Bosco

-0.37

gelöst

-0.36

__":

-0.36

POSITIVE LOGITS

cepan

0.45

 berbentuk

0.44

indrical

0.44

 незавершена

0.43

 ostavi

0.41

abestanden

0.41

 rectangular

0.41

StoryboardSegue

0.41

 Савезне

0.41

形状

0.40

Activations Density 0.096%

specific tools or devices and their descriptions

describing shape and structure

No Comments

No Known Activations