INDEX

Explanations

phrases indicating alignment or agreement with certain principles, values, or norms

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate conformity or alignment with policies, values, or standards

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

gins

-0.73

enaries

-0.68

roma

-0.67

zon

-0.66

ikes

-0.63

bodied

-0.63

reau

-0.63

eware

-0.62

erers

-0.61

verage

-0.59

POSITIVE LOGITS

 tradition

1.05

 expectations

0.93

 prevailing

0.92

 precedent

0.92

 norms

0.90

 tenets

0.87

 principles

0.87

 recommendations

0.86

 orthodoxy

0.83

 reality

0.83

Activations Density 0.144%

phrases indicating alignment or agreement with certain principles, values, or norms

phrases that indicate conformity or alignment with policies, values, or standards

No Comments

No Known Activations