INDEX

Explanations

sequential steps in cooking instructions

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

period or before followed by word

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

Negative Logits

geh

-0.41

keli

-0.40

ModelAttribute

-0.39

 Capit

-0.39

 patri

-0.36

 abroad

-0.36

"`

-0.36

Stell

-0.35

DispatchToProps

-0.35

LLI

-0.35

POSITIVE LOGITS

parsedMessage

0.54

antaranya

0.52

tagHelperRunner

0.50

RegressionTest

0.50

GeoNames

0.49

 AssemblyProduct

0.46

SourceChecksum

0.46

PostInfinity

0.45

исленность

0.45

 nahilalakip

0.45

Activations Density 0.025%

sequential steps in cooking instructions

period or before followed by word

No Known Activations