INDEX

Explanations

instances of commentary and evaluative phrases in conversation

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

professional context

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

 beginnetje

-0.49

 Савезне

-0.46

 cherchés

-0.45

Enllaces

-0.41

లాలు

-0.41

tså

-0.40

 муніципалі

-0.40

 estekak

-0.40

formik

-0.39

dostęp

-0.39

POSITIVE LOGITS

 betweenstory

0.52

 calendriers

0.50

ſammen

0.46

InitVars

0.43

ativeness

0.43

TAGE

0.42

 prostitutes

0.41

diye

0.41

.*")]

0.40

rítica

0.40

Activations Density 0.094%

instances of commentary and evaluative phrases in conversation

professional context

No Comments

No Known Activations