INDEX

Explanations

references to the word "who" and its variants in different contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 5-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_5/width_16k/average_l0_68

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.5.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

 experimenta

-0.61

 Brin

-0.59

 grosse

-0.57

 Optimum

-0.57

ته

-0.56

تمام

-0.56

 experiences

-0.56

Brin

-0.55

ματα

-0.53

Eileen

-0.53

POSITIVE LOGITS

Who

1.69

Who

1.61

who

1.60

who

1.44

WHO

1.37

WHO

1.35

 whom

1.35

 hvem

1.31

Quem

1.31

Кто

1.30

Activations Density 0.058%

references to the word "who" and its variants in different contexts

No Comments

No Known Activations