INDEX

Explanations

references to data structures or computational elements in the context of information processing

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

tilde and following punctuation

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

httphttps

-0.60

TypedDataSet

-0.55

RenderAtEndOf

-0.54

تقاوى

-0.54

aarrggbb

-0.54

webElementXpaths

-0.52

cerpt

-0.52

delwed

-0.50

Obrázky

-0.50

 RemoteException

-0.49

POSITIVE LOGITS

 архивлан

0.40

HomeAsUpEnabled

0.35

Independent

0.30

 wahl

0.28

principalTable

0.28

исленность

0.28

čenje

0.28

ſelves

0.28

slidesToShow

0.28

toprule

0.27

Activations Density 0.214%

references to data structures or computational elements in the context of information processing

tilde and following punctuation

No Comments

No Known Activations