INDEX

Explanations

programming-related directory structures and paths

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

directory paths

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

onOptions

-0.52

 utafitiHapana

-0.50

oa̍t

-0.50

хьтан

-0.48

 Signalez

-0.46

 незавершена

-0.44

WriteTagHelper

-0.43

VertexAttrib

-0.43

PreferredItem

-0.43

 wikipagina

-0.42

POSITIVE LOGITS

 nostru

0.48

LEncoder

0.45

 ActivatedRoute

0.44

 Luxor

0.44

 hắn

0.43

 <<<<<<<<<<<<<<

0.42

 Grossman

0.42

 verantwortlich

0.42

 AttributeSet

0.41

 CommandType

0.39

Activations Density 0.007%

programming-related directory structures and paths

directory paths

No Comments

No Known Activations