INDEX

Explanations

programming structures and elements typically related to function definitions and parameter settings

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

 nakalista

-0.72

riwal

-0.51

 bParam

-0.51

 tartalomajánló

-0.50

 esternos

-0.49

 ComVisible

-0.49

HideFlags

-0.48

expandindo

-0.47

Посилання

-0.47

 setOpen

-0.47

POSITIVE LOGITS

 Eigent

0.37

 ospiti

0.37

 propias

0.37

 geografic

0.36

 propia

0.36

CÁ

0.36

 propios

0.35

 PyTuple

0.35

Diweddarwch

0.34

Rè

0.34

Activations Density 0.203%

programming structures and elements typically related to function definitions and parameter settings

No Comments

No Known Activations