INDEX

Explanations

verbs and adjectives that express participation or activity

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 18-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_18/width_16k/average_l0_74

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.18.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

 للاسماء

-0.80

 wikipagina

-0.75

 marquées

-0.71

 démocratie

-0.70

 CreateTagHelper

-0.70

 lèvres

-0.66

 peggio

-0.65

 sorella

-0.65

 mediodía

-0.64

 nuages

-0.64

POSITIVE LOGITS

tine

0.69

 thing

0.65

 alta

0.64

ELY

0.61

のも

0.59

ly

0.59

 ones

0.58

 sự

0.58

ViewGroup

0.58

 entity

0.57

Activations Density 0.783%

verbs and adjectives that express participation or activity

No Comments

No Known Activations