INDEX

Explanations

occurrences of programming-related structures and syntax

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

apache and android

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

principalColumn

-0.61

 esternos

-0.59

HostException

-0.56

TagMode

-0.54

rawDesc

-0.50

 Савезне

-0.49

 lenker

-0.48

újo

-0.46

Autoritní

-0.45

tanleria

-0.44

POSITIVE LOGITS

ViewFeatures

0.40

 Covenant

0.39

 Inflammation

0.38

 coven

0.38

dumps

0.38

➋

0.37

WebVitals

0.37

AutoScaleMode

0.37

anganronpa

0.37

 certiorari

0.35

Activations Density 0.014%

occurrences of programming-related structures and syntax

apache and android

No Comments

No Known Activations