INDEX

Explanations

negative markers or prefixes in relation to evaluations or situations

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

proper nouns after separators

np_acts-logits-general · gemini-2.5-flash-lite

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

(!__

-0.40

รร

-0.39

 typelib

-0.38

 oprot

-0.38

po

-0.36

 sumpay

-0.35

TacToe

-0.34

ModelAttribute

-0.34

qtype

-0.33

exitRule

-0.32

POSITIVE LOGITS

 nahilalakip

0.54

 незавершена

0.53

complexContent

0.49

 OFDb

0.48

 OnInit

0.47

Traveling

0.47

 houſe

0.47

 mergeFrom

0.46

 Houſe

0.46

 Masyarakat

0.46

Activations Density 0.003%

negative markers or prefixes in relation to evaluations or situations

proper nouns after separators

No Comments

No Known Activations