INDEX

Explanations

references to specific locations or apartment numbers

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

New Auto-Interp

Configuration

google/gemma-scope-2b-pt-res/layer_12/width_1m/average_l0_107

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

1,048,576

Data Type

float32

Hook Name

blocks.12.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

fjspx

-0.76

Література

-0.65

 виправивши

-0.65

 препратки

-0.65

 Issus

-0.63

 فريبيس

-0.63

 >=",

-0.61

expandindo

-0.61

seamless

-0.60

 reco

-0.60

POSITIVE LOGITS

 apartment

0.68

Apartment

0.64

apartment

0.62

Apt

0.60

 Apartment

0.60

 apartments

0.60

Apartments

0.54

iglie

0.54

reportWebVitals

0.51

 Apartments

0.49

Activations Density 0.001%

references to specific locations or apartment numbers

No Comments

No Known Activations