INDEX

Explanations

programming-related terminology and constructs, particularly involving data structures and algorithms

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

median and reverse calculation

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

 daß

-0.41

헌

-0.40

TemporalType

-0.40

 webbplats

-0.40

vogel

-0.40

 itemprop

-0.40

lapsingToolbar

-0.40

fizierung

-0.40

-0.39

済み

-0.38

POSITIVE LOGITS

msgTypes

0.52

MessageTagHelper

0.50

 للمعارف

0.49

riwal

0.46

scn

0.46

rsiniz

0.45

цездатний

0.44

 برانيه

0.42

aarrggbb

0.41

transQ

0.41

Activations Density 0.002%

programming-related terminology and constructs, particularly involving data structures and algorithms

median and reverse calculation

No Comments

No Known Activations