INDEX

Explanations

names of individuals and their affiliations or titles

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

reporting speech and attribution

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

guruan

-0.38

Referanser

-0.37

enterOuterAlt

-0.35

 lipat

-0.34

errHandler

-0.34

unehmen

-0.34

írus

-0.33

 apparaître

-0.32

createComponent

-0.32

 unie

-0.32

POSITIVE LOGITS

 ControllerBase

0.54

AndEndTag

0.53

Pautan

0.52

principalColumn

0.51

󠁬

0.50

bomba

0.49

forKey

0.48

UpInside

0.47

">—

0.46

bom

0.46

Activations Density 0.111%

names of individuals and their affiliations or titles

reporting speech and attribution

No Comments

No Known Activations