INDEX

Explanations

informal negative contractions and indicators of uncertainty or conditions

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

 AssemblyCulture

-0.52

 voeren

-0.49

 onOptions

-0.49

 caráter

-0.48

aktery

-0.48

geordnet

-0.48

 künftig

-0.48

 seragam

-0.47

aarrggbb

-0.46

 Erscheinung

-0.46

POSITIVE LOGITS

 Administrativna

0.48

:+:

0.46

fire

0.43

 Fire

0.42

 lesson

0.40

 Harsh

0.40

NB

0.40

Incomplete

0.40

Zeta

0.40

Fire

0.39

Activations Density 0.007%

informal negative contractions and indicators of uncertainty or conditions

No Comments

No Known Activations