INDEX

Explanations

phrases indicating priorities or importance

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_20/width_16k/average_l0_71

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

хьтан

-0.69

írus

-0.54

 esternos

-0.54

ItemBackground

-0.52

씩

-0.48

 Paglinawan

-0.48

 Segen

-0.48

 Nachteile

-0.48

Ссылки

-0.47

 gese

-0.47

POSITIVE LOGITS

 priority

2.85

 priorities

2.55

 Priority

2.46

priority

2.46

Priority

2.27

 prioritize

2.22

 Priorities

2.22

Priorities

2.18

 prioritized

2.18

 prioritise

2.13

Activations Density 0.190%

phrases indicating priorities or importance

No Comments

No Known Activations