INDEX

Explanations

keywords indicating significance and relationships within a context

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

value and importance

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

GEBURTSDATUM

-0.68

principalColumn

-0.51

 Infórmanos

-0.49

OCCURRED

-0.45

 ベストアンサー

-0.44

 vermelhas

-0.43

fromnode

-0.42

RTGC

-0.42

 Presidencia

-0.42

 censiti

-0.41

POSITIVE LOGITS

AccessorTable

0.53

 Roskov

0.50

SuspendLayout

0.47

 surla

0.46

)}>

0.45

()}>

0.44

|]

0.44

]&

0.43

extAlignment

0.43

SequentialGroup

0.43

Activations Density 0.168%

keywords indicating significance and relationships within a context

value and importance

No Comments

No Known Activations