INDEX

Explanations

phrases related to suspicion or suspected actions

oai_token-act-pair · gpt-3.5-turbo

the term "suspected" in various contexts, particularly related to allegations or assumptions about illegal or dangerous activities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

ammy

-0.82

ajo

-0.76

learn

-0.74

skill

-0.71

reen

-0.70

learning

-0.68

alach

-0.66

OVA

-0.65

www

-0.63

Scot

-0.63

POSITIVE LOGITS

 suspected

0.90

 suspect

0.87

 suspects

0.80

 culprit

0.77

 accused

0.73

 mishand

0.73

 accuse

0.72

 suspicious

0.67

 arson

0.66

 leaking

0.65

Activations Density 0.006%

phrases related to suspicion or suspected actions

the term "suspected" in various contexts, particularly related to allegations or assumptions about illegal or dangerous activities

No Comments

No Known Activations