INDEX

Explanations

categories and directives related to setting preferences or instructions

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

aarrggbb

-0.38

-0.35

 Type

-0.35

 Schiller

-0.34

pro

-0.34

一

-0.33

 heavily

-0.33

 inter

-0.33

Type

-0.33

çon

-0.32

POSITIVE LOGITS

 kasarigan

0.73

Datuak

0.68

 EconPapers

0.65

ScopeManager

0.65

httphttps

0.63

 незавершена

0.62

তথ্যসূত্র

0.62

Tembelea

0.60

 estekak

0.60

 noDo

0.60

Activations Density 0.000%

categories and directives related to setting preferences or instructions

No Comments

No Known Activations