INDEX

Explanations

words related to preferences

oai_token-act-pair · gpt-3.5-turbo

references to personal preferences and choices

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

mberg

-0.80

mans

-0.76

Adams

-0.75

wordpress

-0.75

WARN

-0.72

bane

-0.72

Sus

-0.70

Interstitial

-0.69

manship

-0.68

icles

-0.67

POSITIVE LOGITS

 preferences

1.09

 favoring

0.95

 preference

0.91

eering

0.87

 pane

0.80

 favoured

0.80

 favors

0.79

ļéĨĴ

0.74

 elig

0.72

 favored

0.72

Activations Density 0.018%

words related to preferences

references to personal preferences and choices

No Comments

No Known Activations