INDEX

Explanations

sequences or patterns reflecting numerical values, particularly in code or data contexts

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

midway

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

CloseOperation

-0.68

 bluzka

-0.57

RTDA

-0.54

ThroughAttribute

-0.51

GEBURTSDATUM

-0.50

AnimationsModule

-0.47

RepeatedField

-0.47

matchCondition

-0.46

MethodManager

-0.46

 للمعارف

-0.45

POSITIVE LOGITS

tagHelperRunner

0.49

uari

0.40

iann

0.39

 common

0.38

siv

0.37

吋

0.36

béco

0.35

%\[

0.35

 multi

0.35

Belle

0.35

Activations Density 0.227%

sequences or patterns reflecting numerical values, particularly in code or data contexts

midway

No Comments

No Known Activations