INDEX

Explanations

negations or phrases indicating prohibition 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

 Conclusions

-1.71

 specificity

-1.58

 applications

-1.58

 conclusions

-1.52

 Help

-1.45

 clinicians

-1.45

 clearance

-1.44

ijing

-1.43

 recommendations

-1.43

 Solution

-1.41

POSITIVE LOGITS

Ī

2.41

ľĵ

2.31

Ŀ

2.11

İ

2.08

Ķ

2.06

Ļª

2.00

Ĵ

1.97

ģ

1.96

½

1.95

£

1.91

Activations Density 0.694%

negations or phrases indicating prohibition or restrictions

No Comments

No Known Activations