INDEX

Explanations

specific types of entities, particularly related to names or identifiers in programming contexts

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

before certain suffixes (nik, ant, ous, etc.)

np_acts-logits-general · gemini-2.0-flash

word completion

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

oa̍t

-0.75

 betweenstory

-0.63

 Paulista

-0.61

 мәкалә

-0.57

expandindo

-0.57

 Numerade

-0.57

 Италијани

-0.56

SequentialGroup

-0.56

 resourceCulture

-0.55

epam

-0.55

POSITIVE LOGITS

 pleaſure

0.52

ſelf

0.52

 iſt

0.46

 pngtree

0.46

neſs

0.45

 itſelf

0.44

 Touchable

0.44

 ſta

0.42

 deſt

0.41

 raiſ

0.41

Activations Density 0.470%

specific types of entities, particularly related to names or identifiers in programming contexts

before certain suffixes (nik, ant, ous, etc.)

word completion

No Comments

No Known Activations