INDEX

Explanations

comparative or superlative adjectives indicating levels or degrees

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

New Auto-Interp

Configuration

google/gemma-scope-9b-pt-mlp/layer_6/width_131k/average_l0_72

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.6.hook_mlp_out

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 оригіналу

-0.42

 keen

-0.41

RTLR

-0.41

agd

-0.39

 AssemblyCulture

-0.38

 <>",

-0.37

 Tiefen

-0.37

 насељу

-0.36

 flesta

-0.36

falda

-0.35

POSITIVE LOGITS

#+#

0.52

 becauſe

0.49

extAlignment

0.47

 houſe

0.45

Бахар

0.44

jelasan

0.42

niająca

0.42

__*/

0.42

 defaultstate

0.42

ſelf

0.41

Activations Density 0.005%

comparative or superlative adjectives indicating levels or degrees

No Comments

No Known Activations