INDEX

Explanations

references to macro-level concepts and terminologies

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

providedIn

-0.43

-0.41

WriteLiteral

-0.40

fing

-0.40

 Ewing

-0.40

setColumns

-0.38

Wing

-0.38

ifs

-0.38

saida

-0.38

Ell

-0.37

POSITIVE LOGITS

 Macro

1.11

 macro

1.06

Macro

1.03

 MACRO

0.95

 macros

0.90

macro

0.89

 makro

0.88

 Macros

0.87

MACRO

0.87

 macroeconomic

0.83

Activations Density 0.004%

references to macro-level concepts and terminologies

No Comments

No Known Activations