INDEX

Explanations

phrases related to taking action or actively doing something

oai_token-act-pair · gpt-3.5-turbo

phrases related to acting or being in a role

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

 Gloss

-0.66

artifacts

-0.65

brew

-0.65

aways

-0.65

 sinks

-0.64

 Printed

-0.64

comings

-0.62

arnaev

-0.62

 Riding

-0.61

merce

-0.61

POSITIVE LOGITS

 opposite

0.84

 behalf

0.79

 differently

0.77

 metic

0.75

RECT

0.72

athetic

0.68

uously

0.68

 actionGroup

0.68

 unilaterally

0.68

umbn

0.67

Activations Density 0.094%

phrases related to taking action or actively doing something

phrases related to acting or being in a role

No Comments

No Known Activations