INDEX

Explanations

code-related terms and structures in programming or technical documents

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

New Auto-Interp

Configuration

google/gemma-scope-9b-pt-res/layer_20/width_131k/average_l0_114

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 betweenstory

-0.63

 AssemblyVersion

-0.63

bootstrapcdn

-0.56

Spoljašnje

-0.55

Datuak

-0.55

CodedInputStream

-0.54

 PeEnEo

-0.52

jooq

-0.51

varargin

-0.50

rbrakk

-0.50

POSITIVE LOGITS

MemoryWarning

0.33

leroi

0.33

 morals

0.32

 Kette

0.29

 pinulongan

0.29

 vật

0.29

 flying

0.28

Matriz

0.28

 Indigo

0.28

 clubhouse

0.28

Activations Density 0.026%

code-related terms and structures in programming or technical documents

No Comments

No Known Activations