INDEX

Explanations

references to specific scientific terms or variable notations in a technical context

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

citations \[[.]\]

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

 useAppContext

-0.56

Przypisy

-0.52

Viited

-0.52

Personendaten

-0.51

IContainer

-0.48

исленность

-0.45

CloseOperation

-0.44

OVERR

-0.44

 odkazy

-0.43

ंदीखरीदारी

-0.43

POSITIVE LOGITS

TokenNameLPAREN

0.52

⁩

0.47

 isNameExpr

0.45

utives

0.45

jenigen

0.44

Rhestr

0.44

̣i

0.42

 addirittura

0.41

mTitle

0.41

 يتيمه

0.41

Activations Density 0.025%

references to specific scientific terms or variable notations in a technical context

citations \[[.]\]

No Comments

No Known Activations