INDEX

Explanations

elements related to puzzles and problem-solving

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

pattern of puzzles

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

 reserve

-0.31

 прош

-0.31

 reserved

-0.30

 Chrif

-0.28

 Mika

-0.28

\{\\

-0.28

rasz

-0.28

 vastaan

-0.28

 Frühling

-0.28

 fringe

-0.27

POSITIVE LOGITS

 AssemblyCulture

0.80

 disambiguazione

0.76

featureID

0.75

 виправивши

0.74



0.69

 OFDb

0.68

RTEE

0.67

httphttps

0.65

 FetchType

0.64

 EClass

0.63

Activations Density 0.082%

elements related to puzzles and problem-solving

pattern of puzzles

No Comments

No Known Activations