INDEX

Explanations

the word 'trust'

oai_token-act-pair · gpt-3.5-turbo

references to the concept of trust

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

pmwiki

-0.92

atre

-0.84

issance

-0.71

ozo

-0.71

ploy

-0.71

 Vide

-0.70

neapolis

-0.70

nesota

-0.69

 Invention

-0.69

OPLE

-0.67

POSITIVE LOGITS

worthiness

1.65

lessly

0.98

worthy

0.97

ees

0.91

 trust

0.90

fulness

0.86

fully

0.81

 trusts

0.79

 trusting

0.79

lessness

0.74

Activations Density 0.015%

the word 'trust'

references to the concept of trust

No Comments

No Known Activations