INDEX

Explanations

statements that indicate truths or valid assertions in a technical context

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

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

sizeCache

-0.84

 betweenstory

-0.83

 autorytatywna

-0.82

Tembelea

-0.79

webElementXpaths

-0.79

writeFieldEnd

-0.77

 Roskov

-0.76

 EconPapers

-0.74

 Taktlose

-0.73

تقاوى

-0.73

POSITIVE LOGITS

 exklu

0.35

 Cataluña

0.33

进

0.31

 miroir

0.30

 technically

0.30

 Wagner

0.30

vœ

0.29

 coincid

0.29

 exclusivo

0.29

 Putri

0.29

Activations Density 0.019%

statements that indicate truths or valid assertions in a technical context

No Comments

No Known Activations