INDEX

Explanations

programming-related output statements and function terminators

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

programming output statements

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

AndEndTag

-0.72

 EconPapers

-0.66

 JpaRepository

-0.61

IndentedString

-0.56

tagHelperRunner

-0.56

thâu

-0.55

httphttps

-0.54

yym

-0.54

uxxxx

-0.51

 &___

-0.50

POSITIVE LOGITS

WaitGroup

0.47

MenuInflater

0.41

 AssemblyProduct

0.34

vailability

0.32

AutoresizingMask

0.32

MOL

0.32

fjspx

0.31

 poch

0.31

toid

0.31

 Allahabad

0.31

Activations Density 0.003%

programming-related output statements and function terminators

programming output statements

No Comments

No Known Activations