INDEX

Explanations

references to abstract concepts and values

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

phrases of justification or explanation

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

 tuong

-0.45

StoryboardSegue

-0.40

hadiran

-0.40

 clicar

-0.39

 typelib

-0.37

armi

-0.37

こそ

-0.36

stateMutability

-0.36

GEBURTS

-0.36

 Ausblick

-0.36

POSITIVE LOGITS

:✨

0.47

posedge

0.46

 CreateTagHelper

0.42

 disambiguazione

0.41

ByVersion

0.41

Упаковка

0.39

 caprice

0.38

 bazar

0.38

 disgrace

0.38

ьаж

0.37

Activations Density 0.139%

references to abstract concepts and values

phrases of justification or explanation

No Comments

No Known Activations