INDEX

Explanations

programming-related syntax and structure

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

programming identifiers and names

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

 nahilalakip

-0.55

DockStyle

-0.49

AnchorStyles

-0.40

abestanden

-0.35

RunWith

-0.35

TagMode

-0.34

RTLD

-0.33

menea

-0.33

tonode

-0.33

InjectAttribute

-0.33

POSITIVE LOGITS

 '\\;'

0.63

 iprot

0.59

 يتيمه

0.56

 ModelRenderer

0.55

CppMethod

0.54

 intStringLen

0.54

ftagPool

0.53

 nonUne

0.51

 snippetHide

0.50

withIdentifier

0.47

Activations Density 0.141%

programming-related syntax and structure

programming identifiers and names

No Comments

No Known Activations