INDEX

Explanations

words related to predictions or signs indicating future events

oai_token-act-pair · gpt-3.5-turbo

references to forecasting or predictions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

BuyableInstoreAndOnline

-0.87

RED

-0.77

REDACTED

-0.77

å°Ĩ

-0.70

Downloadha

-0.68

 Norn

-0.68

IRO

-0.68

 Stain

-0.67

 Shed

-0.66

channelAvailability

-0.65

POSITIVE LOGITS

nsics

1.10

told

1.06

shadow

1.06

sight

1.04

runner

1.00

shore

0.99

father

0.98

closed

0.98

warn

0.95

thought

0.94

Activations Density 0.011%

words related to predictions or signs indicating future events

references to forecasting or predictions

No Comments

No Known Activations