INDEX

Explanations

phrases questioning or challenging commonly held beliefs or claims

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

claims and statements regarding beliefs

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

Negative Logits

yntaxException

-0.55

 CreateTagHelper

-0.54

 encomp

-0.49

 AssemblyTitle

-0.48

 consecuencia

-0.46

 typelib

-0.46

vgili

-0.46

FXML

-0.45

aarrggbb

-0.45

 probablemente

-0.45

POSITIVE LOGITS

RTDA

0.45

 везе

0.36

hyrchwyd

0.36

gc

0.35

webElementXpaths

0.33

 بيها

0.33

*/].

0.32

 Malo

0.31

optionalTypeArgs

0.31

 claimed

0.31

Activations Density 0.056%

phrases questioning or challenging commonly held beliefs or claims

claims and statements regarding beliefs

No Known Activations