INDEX

Explanations

references to one's current position or state in relation to goals or progress

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

current position or state

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

 pluie

-0.33

 acrylique

-0.32

 trade

-0.31

vyu

-0.31

 škole

-0.31

rozco

-0.30

 cesse

-0.30

fiez

-0.30

 hâte

-0.30

 setPrice

-0.29

POSITIVE LOGITS

InitVars

0.63

 milestone

0.62

 standstill

0.61

Tembelea

0.61

AlterField

0.60

 reached

0.60

SharedCtor

0.57

reached

0.57

 impasse

0.57

principalTable

0.56

Activations Density 0.044%

references to one's current position or state in relation to goals or progress

current position or state

No Comments

No Known Activations