INDEX

Explanations

names or terms related to individuals, particularly with the syllable "ong" or variations thereof

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

names ending in ong

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

*</

-0.54

))*

-0.50

readAs

-0.49

])*

-0.49

 createSlice

-0.44

 fallu

-0.43

Tea

-0.42

 getField

-0.42

*}

-0.42

POSITIVE LOGITS

ong

1.24

ONG

1.06

Ong

0.96

ongs

0.92

Ong

0.91

ong

0.90

thong

0.90

 Thong

0.89

orong

0.88

onger

0.86

Activations Density 0.014%

names or terms related to individuals, particularly with the syllable "ong" or variations thereof

names ending in ong

No Comments

No Known Activations