INDEX

Explanations

code-related structures and expressions in programming languages

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

WriteTagHelper

-1.02

parsedMessage

-0.98

 NSCoder

-0.94

GEBURTSDATUM

-0.94



-0.91

bootstrapcdn

-0.88

 فريبيس

-0.85

EDEFAULT

-0.84

AddTagHelper

-0.82

 ویکی‌پدی

-0.77

POSITIVE LOGITS

 works

0.27

Pid

0.23

0.22

it

0.22

run

0.21

 horizon

0.21

đ

0.21

 rain

0.21

 using

0.21

Activations Density 0.011%

code-related structures and expressions in programming languages

No Comments

No Known Activations