INDEX

Explanations

active verbs and phrases indicating physical or dynamic actions

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

repetition or recurrence

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

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 informée

-0.52

utnik

-0.50

AGGIO

-0.45

jspb

-0.45

GetKeyDown

-0.43

NoteOnEvent

-0.43

 DataLoader

-0.43

 ObjectMapper

-0.42

 LatLng

-0.42

glLoad

-0.41

POSITIVE LOGITS

 again

0.91

again

0.78

Again

0.69

 AGAIN

0.67

 Again

0.67

 novamente

0.64

 nuevamente

0.61

 다시

0.60

 lagi

0.60

 nuovamente

0.57

Activations Density 0.018%

active verbs and phrases indicating physical or dynamic actions

repetition or recurrence

No Comments

No Known Activations