INDEX

Explanations

words that indicate prevalence or frequency

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

 qualsiasi

-0.65

 qualunque

-0.63

 viață

-0.61

 habet

-0.60

جوايز

-0.60

⋂

-0.59

 chamomile

-0.58

 syke

-0.57

 încă

-0.57

 måned

-0.56

POSITIVE LOGITS

 Mostly

2.05

Mostly

2.03

 mainly

2.03

 Mainly

2.00

 mostly

1.97

Mainly

1.88

mostly

1.86

mainly

1.83

 chiefly

1.75

 primarily

1.74

Activations Density 0.151%

words that indicate prevalence or frequency

No Comments

No Known Activations