INDEX

Explanations

phrases involving personal pronouns and their antecedents, particularly focusing on "who" and "which" clauses

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 9-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_9/width_16k/average_l0_73

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.9.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

 dommage

-0.50

 aceste

-0.47

Semi

-0.46

 dovol

-0.46

 disponibilités

-0.46

ρω

-0.46

 fondament

-0.45

 najbolj

-0.45

раздо

-0.45

 rigtig

-0.45

POSITIVE LOGITS

 οποίο

0.90

SequentialGroup

0.77

">—

0.75

 also

0.73

 οποία

0.72

ίος

0.70

UMBIA

0.70

 שוליים

0.70

 autorytatywna

0.69

__':

0.68

Activations Density 0.298%

phrases involving personal pronouns and their antecedents, particularly focusing on "who" and "which" clauses

No Comments

No Known Activations