INDEX

Explanations

words and phrases that indicate contrasting opinions or outcomes

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

 Overrides

-0.48

outi

-0.42

izedBox

-0.42

logenetic

-0.42

klinik

-0.42

XmlAccessType

-0.42

 rodillas

-0.41

⅘

-0.41

 Dram

-0.41

ementara

-0.41

POSITIVE LOGITS

CppMethod

0.49

 cherchés

0.45

 EconPapers

0.38

Datuak

0.35

 együtt

0.33

akei

0.32

httphttps

0.32

======

0.32

 comuniques

0.32

usercontent

0.32

Activations Density 0.043%

words and phrases that indicate contrasting opinions or outcomes

No Comments

No Known Activations