INDEX

Explanations

expressions that represent function calls or method invocations

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

email address or code variables

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

AddTagHelper

-0.60

Tikang

-0.56

 invokingState

-0.54

 dAtA

-0.50

 للاسماء

-0.48

RTEE

-0.48

orologio

-0.47

 hacerte

-0.47

UnusedPrivate

-0.47

ForRow

-0.47

POSITIVE LOGITS

CGRectMake

0.45

oa̍t

0.42

BeginContext

0.42

 كومونز

0.42

 transférez

0.41

Спасылкі

0.39

amerikanischer

0.36

[]}

0.36

TestBed

0.35

 EconPapers

0.35

Activations Density 0.030%

expressions that represent function calls or method invocations

email address or code variables

No Comments

No Known Activations