INDEX

Explanations

punctuations and symbols associated with written text

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

programming code identifiers

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

DOF

-0.49

ubernur

-0.48

andowski

-0.44

amoan

-0.42

៛

-0.42

VOA

-0.42

cipar

-0.41

 indígen

-0.41

Prove

-0.41

 recor

-0.41

POSITIVE LOGITS

 AssemblyTitle

0.64

RenderAtEndOf

0.47

 onCreateView

0.47

ParallelGroup

0.46

MemoryWarning

0.45

 referenties

0.45

UnusedPrivate

0.45

LookAnd

0.44

 ChromeDriver

0.44

writeFieldEnd

0.44

Activations Density 0.014%

punctuations and symbols associated with written text

programming code identifiers

No Comments

No Known Activations