INDEX

Explanations

references to examples and their outcomes in instructional or procedural texts

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

 smartest

-0.44

 emancip

-0.43

 faſt

-0.42

 noDo

-0.40

ècie

-0.39

İstinadlar

-0.38

 kitten

-0.38

 worst

-0.38

qrstuvwxyz

-0.38

 opium

-0.38

POSITIVE LOGITS

 NSCoder

0.46

 fabbrica

0.45

ETHOD

0.44

aktery

0.44

 EXAMPLE

0.42

 example

0.42

empio

0.41

lexikon

0.41

 Ejemplo

0.41

 esempio

0.41

Activations Density 0.048%

references to examples and their outcomes in instructional or procedural texts

No Comments

No Known Activations