INDEX

Explanations

code structures and definitions in programming

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

defining categories and names

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

IsMutable

-0.58

脚注の使い方

-0.49

FunctionFlags

-0.48

󠁴

-0.46

[];

-0.44

IntoConstraints

-0.44

jsxFileName

-0.44

CROSSTALK

-0.44

intios

-0.43

__':

-0.43

POSITIVE LOGITS

 overall

0.61

 Overall

0.52

Geplaatst

0.51

Overall

0.46

整體

0.45

overall

0.44

tagHelperRunner

0.42

utilisons

0.42

 compilation

0.41

 keseluruhan

0.41

Activations Density 0.671%

code structures and definitions in programming

defining categories and names

No Comments

No Known Activations