INDEX

Explanations

syntactic structures and elements used in programming or markup languages

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

programming keywords and 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

iſen

-0.58

endphp

-0.57

liceerd

-0.56

iſten

-0.56

riwal

-0.55

majánló

-0.54



-0.54

ambién

-0.54

 corrupción

-0.53

 Taktlose

-0.53

POSITIVE LOGITS

AddTagHelper

0.33

principalColumn

0.30

NamedQueries

0.30

resizingMask

0.29

OGND

0.29

addCriterion

0.28

ich

0.28

 ویکی‌پدیای

0.28

Hip

0.27

новить

0.27

Activations Density 0.116%

syntactic structures and elements used in programming or markup languages

programming keywords and identifiers

No Comments

No Known Activations