INDEX

Explanations

elements of code related to function definitions and structure

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

New Auto-Interp

Configuration

google/gemma-scope-9b-pt-res/layer_20/width_131k/average_l0_114

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.20.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

<unused8>

-0.91

<unused52>

-0.91

parsedMessage

-0.90

<unused47>

-0.90

<unused79>

-0.90

<unused41>

-0.90

<unused43>

-0.90

<unused28>

-0.90

[@BOS@]

-0.90

<unused17>

-0.90

POSITIVE LOGITS

 pañ

0.27

 likewise

0.27

 vuel

0.27

 équi

0.25

 poème

0.25

 desnuda

0.25

 Wohnzimmer

0.24

 afuera

0.24

 also

0.24

 apont

0.24

Activations Density 0.012%

elements of code related to function definitions and structure

No Comments

No Known Activations