INDEX

Explanations

key terms related to research findings and challenges in experimentation

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

 sacré

-0.39

 čím

-0.34

 stabilisation

-0.33

ɥ

-0.33

 loveliness

-0.33

 dammit

-0.32

 civilised

-0.32

 मैंने

-0.31

 savoury

-0.31

 darn

-0.31

POSITIVE LOGITS

onViewCreated

0.74

SharedDtor

0.65

+#+#

0.56

transQ

0.56

 podjela

0.56

fortawesome

0.54

posedge

0.53

abestanden

0.50

GIVEREF

0.50

 ſind

0.49

Activations Density 0.320%

key terms related to research findings and challenges in experimentation

No Comments

No Known Activations