INDEX

Explanations

references to programming constructs, particularly packages and modules

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

Code snippets, often file paths or namespaces

np_acts-logits-general · gemini-2.0-flash

code objects and addons

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-16k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

SequentialGroup

-0.59

AndEndTag

-0.47

astéro

-0.42

<bos>

-0.41

 propOrder

-0.41

comuna

-0.40

picasso

-0.40

 nakalista

-0.40

__":

-0.39

+:+

-0.39

POSITIVE LOGITS

endcsname

0.50

rungsseite

0.49

 <<<<<<<<<<<<<<

0.48

AppRoutingModule

0.44

kj

0.42

 صوتيه

0.41

μφωνα

0.41

 Anglia

0.40

ud

0.40

Rohy

0.40

Activations Density 0.145%

references to programming constructs, particularly packages and modules

Code snippets, often file paths or namespaces

code objects and addons

No Comments

No Known Activations