INDEX

Explanations

phrases indicating capability or potential, particularly verbs associated with possibility and action

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 21-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_21/width_16k/average_l0_129

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.21.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 tend

-0.35

-0.34

don

-0.33

-0.31

’

-0.30

-0.29

es

-0.28

 chercher

-0.28

opt

-0.28

POSITIVE LOGITS

 canst

0.83

 queſta

0.76

extAlignment

0.73

verwijspagina

0.72

ſelves

0.69

 pouvoit

0.69

<unused52>

0.67

<unused21>

0.67

<unused47>

0.67

<unused68>

0.67

Activations Density 0.878%

phrases indicating capability or potential, particularly verbs associated with possibility and action

No Comments

No Known Activations