INDEX

Explanations

programming terms related to variables and methods used in code

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

initial or previous state

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

 EconPapers

-0.60

 queſta

-0.58

AndEndTag

-0.51

sprzedam

-0.51

 Infór

-0.50

 Vordergrund

-0.50

 meille

-0.49

 mukana

-0.48

 ſind

-0.48

 amarillas

-0.47

POSITIVE LOGITS

Demografie

0.39

 NSCoder

0.39

Rie

0.34

tup

0.33

时的

0.33

Rig

0.33

writeFieldEnd

0.32

WithIOException

0.31

ElementException

0.31

Ped

0.31

Activations Density 0.008%

programming terms related to variables and methods used in code

initial or previous state

No Comments

No Known Activations