INDEX

Explanations

numerical values and symbols used in code or data structures

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

code syntax, punctuation

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

⟬

-0.48

oty

-0.47

 EconPapers

-0.43

MockMvc

-0.43

soy

-0.42

 nahilalakip

-0.42

ultuous

-0.42

tonsoft

-0.42

Autoritní

-0.41

 Princes

-0.41

POSITIVE LOGITS

 případ

0.47



0.45

 potenciales

0.44

 alimentaria

0.43

:✨

0.43

 թվական

0.40

 separado

0.40

 hormigón

0.40

 szóci

0.39

 kiin

0.38

Activations Density 0.038%

numerical values and symbols used in code or data structures

code syntax, punctuation

No Comments

No Known Activations