INDEX

Explanations

words related to verbal expression or declarations

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

 Conc

-0.63

setopt

-0.62

vic

-0.62

BeginContext

-0.62

FED

-0.61

vip

-0.60

VIP

-0.60

 Hadid

-0.57

BIC

-0.57

VAC

-0.57

POSITIVE LOGITS

utter

0.66

 literario

0.63

 monasterio

0.61

 colgantes

0.59

 inoxidable

0.57

 Flutter

0.56

 poesía

0.56

 meditación

0.56

 deportivas

0.55

 embarazadas

0.55

Activations Density 0.013%

words related to verbal expression or declarations

No Comments

No Known Activations