INDEX

Explanations

terms related to class definitions and their attributes in programming structures

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

 oprot

-0.60

 Geſch

-0.60

 transfieras

-0.57

DockStyle

-0.55

 Exactos

-0.52

 Мексичка

-0.51

melten

-0.50

InputBorder

-0.50

Eup

-0.49

EndContext

-0.49

POSITIVE LOGITS

delwed

0.44

 feature

0.38

 special

0.33

providedIn

0.33

features

0.33

 özel

0.33

ftagPool

0.33

 features

0.33

feature

0.32

useState

0.32

Activations Density 0.173%

terms related to class definitions and their attributes in programming structures

No Comments

No Known Activations