INDEX

Explanations

prepositions and articles

oai_token-act-pair · gpt-3.5-turbo

phrases indicating actions and relationships in contexts involving significant events or topics

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 toget

-0.77

 chloride

-0.74

 sidew

-0.73

>[

-0.73

ylene

-0.72

anton

-0.72

agra

-0.71

 dispos

-0.70

 culp

-0.70

execute

-0.68

POSITIVE LOGITS

 Excellence

1.06

 Homeless

0.97

 Month

0.96

 Lights

0.95

 Seconds

0.94

 Hour

0.94

 Bullets

0.94

 Monkey

0.93

Hyp

0.93

 Billion

0.93

Activations Density 0.383%

prepositions and articles

phrases indicating actions and relationships in contexts involving significant events or topics

No Comments

No Known Activations