INDEX

Explanations

references to groups of people, particularly using casual address terms

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

 Fortun

-0.61

 defire

-0.61

 retirement

-0.60

 Apare

-0.60

 gammel

-0.59

 adultery

-0.59

 autobiography

-0.59

 Chet

-0.58

 industriels

-0.58

 criança

-0.57

POSITIVE LOGITS

Gentlemen

0.76

 >=",

0.70

 yourselves

0.66

 Audiodateien

0.65

 yall

0.64

бята

0.64

ConfigureAwait

0.62

rawDesc

0.62

ؤلاء

0.60

 guys

0.60

Activations Density 0.088%

references to groups of people, particularly using casual address terms

No Comments

No Known Activations