INDEX

Explanations

terms related to programming and data structures

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

followed by prepositions or punctuation

np_acts-logits-general · gemini-2.0-flash

appointed, elected, clip, torn

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-16k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

jspb

-0.50

 Vaughan

-0.49

errHandler

-0.45

iprot

-0.44

 restricted

-0.44

toHaveBeenCalled

-0.43

PreExecute

-0.41

getInputStream

-0.41

новниш

-0.41

setAction

-0.40

POSITIVE LOGITS

 NSCoder

0.65

 plutôt

0.50

instead

0.47

rrggbb

0.46

rather

0.45

ConstraintMaker

0.43

 navideñas

0.42

 instead

0.42

 nakalista

0.42

 rather

0.42

Activations Density 1.002%

terms related to programming and data structures

followed by prepositions or punctuation

appointed, elected, clip, torn

No Comments

No Known Activations