INDEX

Explanations

the use of the verb "do" in various forms to indicate actions or activities

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

did not / does not

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

ConstraintMaker

-0.75

rungsseite

-0.62

⟬

-0.53

jsonwebtoken

-0.52

 Chwiliwch

-0.52

bootstrapcdn

-0.51

verwijspagina

-0.51

GTCX

-0.50

WithIOException

-0.49

htra

-0.49

POSITIVE LOGITS

 étoit

0.55

 feroit

0.51

eseorang

0.47

 ferien

0.45

ulitan

0.44

 auroit

0.43

 religione

0.43

 cadeia

0.43

 Bulan

0.43

 enfans

0.43

Activations Density 0.054%

the use of the verb "do" in various forms to indicate actions or activities

did not / does not

No Comments

No Known Activations