INDEX

Explanations

various types of symbols and punctuation marks

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

 Савезне

-0.38

 alrededores

-0.29

 préc

-0.28

又不

-0.26

又

-0.25

zarchiwizowane

-0.25

 boîtier

-0.25

DispatchToProps

-0.23

 كمان

-0.23

 Jenderal

-0.23

POSITIVE LOGITS

 AssemblyCulture

0.85

MessageTagHelper

0.66

richTextPanel

0.65

principalTable

0.63



0.62

PostInfinity

0.61

UrlResolution

0.60

Clik

0.60

FormTagHelper

0.59

principalColumn

0.59

Activations Density 0.051%

various types of symbols and punctuation marks

No Comments

No Known Activations