INDEX

Explanations

subscription-related phrases

oai_token-act-pair · gpt-3.5-turbo

phrases related to subscribing to newsletters or email lists

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

 Morg

-0.72

ball

-0.71

bra

-0.70

 Dull

-0.70

 Stephenson

-0.66

 Dahl

-0.65

Gou

-0.64

 Lambert

-0.63

pex

-0.63

 McMaster

-0.63

POSITIVE LOGITS

 subscribe

0.85

 subscribing

0.84

Interstitial

0.82

itte

0.82

 binge

0.80

 subscriptions

0.80

irmation

0.77

 uncond

0.77

essions

0.76

interstitial

0.75

Activations Density 0.011%

subscription-related phrases

phrases related to subscribing to newsletters or email lists

No Comments

No Known Activations