INDEX

Explanations

different forms of negation in the text

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

java

-0.39

 càng

-0.36

liferay

-0.35

this

-0.35

enderror

-0.34

tag

-0.34

yes

-0.33

 indeed

-0.33

RL

-0.32

indeed

-0.32

POSITIVE LOGITS

 pinulongan

0.66

canestro

0.59

 виправивши

0.57

expandindo

0.56

TestingModule

0.56

 nakalista

0.54

 Administrativna

0.54

 NUKAT

0.53

 محفوظة

0.52

Välislingid

0.52

Activations Density 0.045%

different forms of negation in the text

No Comments

No Known Activations