INDEX

Explanations

tokens in mathematical or scientific discourse, like "measure", "using", numbers, and symbols like "g"

oai_token-act-pair · gemini-2.0-flash

scientific terminology

np_token-act-pair-logits · gpt-4o-mini

using

np_max-act-logits · gemini-2.0-flash

New Auto-Interp

Configuration

google/gemma-scope-2b-pt-transcoders/layer_21/width_16k/average_l0_13

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.21.ln2.hook_normalized

Architecture

jumprelu_transcoder

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 myſelf

-1.11

 himſelf

-1.09

 themſelves

-1.01

 fubject

-1.00

 pleaſure

-0.98

 itſelf

-0.98

 Monfieur

-0.96

 purpoſe

-0.96

 deſt

-0.94

 Majefty

-0.93

POSITIVE LOGITS

Gra

0.53

 work

0.52

}",

0.50

0.48

ra

0.48

de

0.48

Sav

0.48

Sha

0.47

the

0.47

re

0.47

Activations Density 2.453%

tokens in mathematical or scientific discourse, like "measure", "using", numbers, and symbols like "g"

scientific terminology

using

No Comments

No Known Activations