INDEX

Explanations

question marks and conversational prompts

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

discussing or explaining "as for"

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

 ब्रेकडाउन

-0.63

 arşivlendi

-0.62

__(/*!

-0.60

StructEnd

-0.57

Билгалдахарш

-0.55

SourceChecksum

-0.52

ValueStyle

-0.51

 BoxDecoration

-0.51

UnknownFieldSet

-0.50

 EconPapers

-0.49

POSITIVE LOGITS

 depends

0.37

については

0.36

に関しては

0.35

andExpect

0.34

지는

0.34

strokeWeight

0.33

Oh

0.33

0.31

dots

0.31

 comunque

0.30

Activations Density 0.092%

question marks and conversational prompts

discussing or explaining "as for"

No Comments

No Known Activations