INDEX

Explanations

negative phrases related to warranties and liabilities

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

New Auto-Interp

Configuration

ctigges/pythia-70m-deduped__res-sm_processed/1-res-sm

Prompts (Dashboard)

32,768 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

32,768

Data Type

torch.float32

Hook Name

blocks.1.hook_resid_post

Hook Layer

Architecture

standard

Context Size

128

Dataset

EleutherAI/the_pile_deduplicated

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

.",

-1.63

igraphy

-1.60

 indoors

-1.58

".

-1.57

.","

-1.55

inqu

-1.51

 outdoors

-1.47

igraph

-1.47

","

-1.45

.](

-1.44

POSITIVE LOGITS

 circumstances

1.74

 conditions

1.60

TRODUCTION

1.56

 parties

1.50

 quotes

1.49

 behalf

1.49

 SUMMARY

1.45

 particular

1.45

 reason

1.43

£

1.43

Activations Density 0.119%

negative phrases related to warranties and liabilities

No Comments

No Known Activations