INDEX

Explanations

representations of different types of data or variable identifiers

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

scientific paper sections or identifiers

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

 arvio

-0.42

 AppCompatTheme

-0.36

újo

-0.35

RTEX

-0.35

 giả

-0.35

DBNull

-0.35

ismet

-0.34

⅘

-0.34

 bestuur

-0.33

 Wikiseite

-0.32

POSITIVE LOGITS

 autorytatywna

0.89

 nahilalakip

0.56

 informée

0.55

rungsseite

0.55

ⓧ

0.52

TagMode

0.51

CppMethod

0.51

findpost

0.50

httphttps

0.50

 незавершена

0.49

Activations Density 0.001%

representations of different types of data or variable identifiers

scientific paper sections or identifiers

No Comments

No Known Activations