INDEX

Explanations

negative constructs in sentences

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

auxiliaries and descriptions

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

AccessorTable

-0.59

principalTable

-0.53

RefNanny

-0.45

NameInMap

-0.42

ngdoc

-0.40

WaitGroup

-0.37

 bezeichneter

-0.36

 nærm

-0.35

InitVars

-0.35

resizingMask

-0.34

POSITIVE LOGITS

LabelTagHelper

0.56

punk

0.52

 polygons

0.51

zuk

0.51

 smartest

0.50

Cup

0.50



0.50

 ſeinen

0.49

 flutes

0.49

 clarinet

0.49

Activations Density 0.038%

negative constructs in sentences

auxiliaries and descriptions

No Comments

No Known Activations