INDEX

Explanations

words related to performing actions or making changes

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

the verbs "make" or "faire"

np_acts-logits-general · gemini-2.0-flash

make clear or faire clear

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_20/width_16k/average_l0_71

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

 trataba

-0.46

 <<<<<<<<<<<<<<

-0.45

make

-0.39

реш

-0.39

ApiOperation

-0.38

ificaciones

-0.37

fattning

-0.36

gotta

-0.36

 Treff

-0.36

 Namely

-0.35

POSITIVE LOGITS

 known

0.87

 available

0.81

 Known

0.78

thâu

0.75

 تضيفلها

0.73

 aware

0.71

 possible

0.70

 KNOWN

0.70

"":

0.68

 evident

0.68

Activations Density 0.128%

words related to performing actions or making changes

the verbs "make" or "faire"

make clear or faire clear

No Comments

No Known Activations