INDEX

Explanations

references to Manhattan and related geographic contexts

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

New Auto-Interp

Configuration

google/gemma-scope-9b-pt-mlp/layer_6/width_131k/average_l0_72

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.6.hook_mlp_out

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

 Manhattan

-1.84

Manhattan

-1.72

 manhattan

-1.44

hattan

-0.73

 Brooklyn

-0.65

 Harlem

-0.61

NYC

-0.60

Brooklyn

-0.59

 brooklyn

-0.59

nyc

-0.54

POSITIVE LOGITS

AndEndTag

0.48

Clik

0.45

 direta

0.43

fvar

0.42

новништво

0.42

pushFollow

0.41

EditorBrowsable

0.40

ourtney

0.39

NullCheck

0.39

 препратки

0.38

Activations Density 0.001%

references to Manhattan and related geographic contexts

No Comments

No Known Activations