INDEX

Explanations

negations and phrases that imply prohibitions or restrictions

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

 simple

-1.60

 clicking

-1.59

 quick

-1.55

ureus

-1.54

 blocking

-1.52

 straightforward

-1.49

irections

-1.46

 pressing

-1.45

 fois

-1.43

ages

-1.43

POSITIVE LOGITS

ħ

4.25

ŀ

4.13

Ŀ

4.12

İ

4.07

ľ

4.05

ģ

4.01

¤

3.97

¾

3.95

Ļ

3.88

Ĩ

3.87

Activations Density 0.132%

negations and phrases that imply prohibitions or restrictions

No Comments

No Known Activations