INDEX

Explanations

various types of closing punctuation and parentheses in code or structured text

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

Python context managers

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 9-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_9/width_131k/average_l0_121

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.9.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

 pinulongan

-0.67

 poffible

-0.54

 nonUne

-0.51

Viited

-0.50

providedIn

-0.49

⟬

-0.48

 purpoſe

-0.47

 beginnetje

-0.46

Plaza

-0.46

IconModule

-0.45

POSITIVE LOGITS

CDATA

0.37

et

0.36

etc

0.35

ніципалі

0.34

 representations

0.34

bibitem

0.34

uxxxx

0.34

 Waray

0.33

 sets

0.33

 experimentally

0.33

Activations Density 0.011%

various types of closing punctuation and parentheses in code or structured text

Python context managers

No Known Activations