INDEX

Explanations

specific numerical patterns or values within data

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

New Auto-Interp

Configuration

ctigges/pythia-70m-deduped__mlp-sm_processed/2-mlp-sm

Prompts (Dashboard)

32,768 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

32,768

Data Type

torch.float32

Hook Name

blocks.2.hook_mlp_out

Hook Layer

Architecture

standard

Context Size

128

Dataset

EleutherAI/the_pile_deduplicated

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

}})$.

-1.82

 behind

-1.78

}})$,

-1.67

'</

-1.65

}})$

-1.65

 cited

-1.57

otted

-1.55

:=

-1.54

 involved

-1.54

(_

-1.53

POSITIVE LOGITS

·

1.54

erals

1.47

quin

1.45

»¿

1.45

¸

1.44

ĻĤ

1.42

¥

1.42

dies

1.42

ser

1.40

APTER

1.39

Activations Density 1.043%

specific numerical patterns or values within data

No Comments

No Known Activations