INDEX

Explanations

code-related syntax and function parameters

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

error codes and parameter attributes

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

 queſta

-0.53

 linkovi

-0.49

VersionUID

-0.48

 autorité

-0.47

 CYTHON

-0.46

partimento

-0.44

انتهای

-0.44

/*

-0.43

 BoxFit

-0.43

 getchar

-0.42

POSITIVE LOGITS

InlineData

0.53

awtextra

0.41

equila

0.40

 Pferde

0.40

lucht

0.39

 olímp

0.39

uncu

0.38

 corrida

0.38

 infecting

0.38

 calculées

0.38

Activations Density 0.027%

code-related syntax and function parameters

error codes and parameter attributes

No Comments

No Known Activations