INDEX

Explanations

words and phrases related to specific types of actions, particularly in the context of describing processes 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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

󠁢

-0.57

 onBind

-0.48

 محفوظة

-0.44

InstanceState

-0.44

fjspx

-0.42

Clik

-0.40

enumi

-0.40



-0.39

ParallelGroup

-0.39

轭

-0.38

POSITIVE LOGITS

Personendaten

0.55

 argint

0.52

ècie

0.46

oggle

0.46

 ProtoMessage

0.45

fromnode

0.44

ByVersion

0.43

 annot

0.42

 dezelve

0.42

hetamine

0.42

Activations Density 0.429%

words and phrases related to specific types of actions, particularly in the context of describing processes or conditions

No Comments

No Known Activations