INDEX

Explanations

elements related to structured data and code representations

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

data delimiters and string

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

findpost

-0.74

ьаж

-0.71

 autorytatywna

-0.69

TagMode

-0.68

AddTagHelper

-0.63

 שוליים

-0.63

<unused41>

-0.61

<unused23>

-0.61

[@BOS@]

-0.61

<unused3>

-0.61

POSITIVE LOGITS

GEBURTS

0.53

fromLTRB

0.35

exh

0.32

<?

0.30

accia

0.30

 Thor

0.30

autoconfigure

0.30

 page

0.29

page

0.28

ced

0.28

Activations Density 0.132%

elements related to structured data and code representations

data delimiters and string

No Comments

No Known Activations