INDEX

Explanations

technical terms related to software development and file management

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

Code, file paths, or directories

np_acts-logits-general · gemini-2.0-flash

folder files

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-16k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

 dead

-0.33

 position

-0.30

нан

-0.30

te

-0.30

day

-0.30

vak

-0.30

 Davis

-0.29

 etme

-0.29

 desk

-0.29

sum

-0.29

POSITIVE LOGITS



0.73

MLLoader

0.71

TypedDataSet

0.69

writeFieldEnd

0.62

 AssemblyVersion

0.58

 pulumi

0.57

 kasarigan

0.57

meteries

0.55

AndEndTag

0.54

parsedMessage

0.54

Activations Density 0.847%

technical terms related to software development and file management

Code, file paths, or directories

folder files

No Comments

No Known Activations