INDEX

Explanations

references to competitive situations and events involving teams or participants

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

teams, representation, Indian, entry

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

 shower

-0.39

 room

-0.38

 surface

-0.37

bou

-0.37

 lune

-0.35

 onset

-0.35

 absorber

-0.34

стры

-0.34

isAuth

-0.34

 interior

-0.34

POSITIVE LOGITS

WaitGroup

0.59

CodeAttribute

0.52

protoimpl

0.50

raszam

0.48

">—

0.48

anmoins

0.45

новниш

0.44

WriteAttribute

0.43

Datuak

0.42

 grossen

0.42

Activations Density 0.015%

references to competitive situations and events involving teams or participants

teams, representation, Indian, entry

No Comments

No Known Activations