INDEX

Explanations

numerical values and potential conditional references

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 18-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_18/width_16k/average_l0_74

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.18.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

aarrggbb

-0.74

:✨

-0.68

ValueStyle

-0.67

 otomatig

-0.62

彿

-0.60

mergeFrom

-0.57

omock

-0.57

ChatColor

-0.56

Referències

-0.56

parsedMessage

-0.56

POSITIVE LOGITS

Personendaten

0.69

awtextra

0.65

parametrize

0.54

nder

0.53

 étudi

0.52

 ruban

0.51

 épisode

0.51

étape

0.50

 Suff

0.50

 Project

0.48

Activations Density 0.327%

numerical values and potential conditional references

No Comments

No Known Activations