INDEX

Explanations

assertions or statements of fact

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

Follows discourse markers (indeed, in fact)

np_acts-logits-general · gemini-2.0-flash

in fact indeed

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

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 obviamente

-0.59

tså

-0.58

 ovviamente

-0.57

 obviously

-0.56

 Obviously

-0.56

จึง

-0.55

TileEntity

-0.55

 respectively

-0.53

afficheront

-0.53

 Федерации

-0.52

POSITIVE LOGITS

 quite

0.81

 even

0.74

 sometimes

0.73

quite

0.69

 meeste

0.69

 might

0.67

 sogar

0.67

Quite

0.65

 plupart

0.65

 zelfs

0.65

Activations Density 0.116%

assertions or statements of fact

Follows discourse markers (indeed, in fact)

in fact indeed

No Comments

No Known Activations