INDEX

Explanations

information related to specific events and personal details, potentially related to investigative work or news reporting

oai_token-act-pair · gpt-3.5-turbo

references to significant events or activities involving individuals

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

 superiority

-0.67

 sacrific

-0.66

 equality

-0.66

 cures

-0.65

iour

-0.65

Improved

-0.65

 optim

-0.64

oppable

-0.64

itiveness

-0.63

 ç¥ŀ

-0.62

POSITIVE LOGITS

 aliases

1.01

 redacted

1.00

 filings

0.98

 LinkedIn

0.94

 Manafort

0.88

 subpoen

0.85

 allegedly

0.84

 iCloud

0.84

 FOIA

0.84

 mailing

0.83

Activations Density 1.364%

information related to specific events and personal details, potentially related to investigative work or news reporting

references to significant events or activities involving individuals

No Comments

No Known Activations