INDEX

Explanations

coding-related terms and concepts within a technical context

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

 nameof

-0.38

Kn

-0.35

tagHelper

-0.34

 https

-0.33

 Dark

-0.32

getKey

-0.32

 pale

-0.31

"#"

-0.30

 zask

-0.30

 knuckle

-0.29

POSITIVE LOGITS

 kasarigan

0.74

findpost

0.62

 uLocal

0.60

TestingModule

0.60

InstrumentedTest

0.59

IsMutable

0.59

 Initialized

0.59

VIRGO

0.59

 Photocase

0.59

 queſta

0.58

Activations Density 0.184%

coding-related terms and concepts within a technical context

No Comments

No Known Activations