INDEX

Explanations

terms related to technical or grammatical structures in language

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Felsen

-0.62

 Decke

-0.61

 Rektor

-0.61

 Kapelle

-0.59

 Ärz

-0.59

 selatan

-0.59

 Autorin

-0.59

 Kirche

-0.58

 Heizung

-0.57

 Arzt

-0.56

POSITIVE LOGITS

 Réponses

0.52

ⓧ

0.49

Bioaccumulative

0.47

!*\

0.45

馃

0.44

prüche

0.43

Consulting

0.42

 Pristupljeno

0.42

RenderAtEndOf

0.41

afp

0.41

Activations Density 0.004%

terms related to technical or grammatical structures in language

No Comments

No Known Activations