INDEX

Explanations

keywords related to sports such as "ickets" and "wicket"

oai_token-act-pair · gpt-3.5-turbo

mentions of tickets or ticketing information

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 centered

-0.71

 sham

-0.70

hol

-0.70

 alpha

-0.68

 contrast

-0.68

 ethics

-0.67

mon

-0.67

 Alpha

-0.65

 sans

-0.64

 Focus

-0.64

POSITIVE LOGITS

ickets

4.90

icket

4.02

Tickets

1.27

 cricket

1.19

 Tickets

1.12

igers

1.12

akedown

1.04

 Lanka

1.02

 Cricket

1.01

ioxide

1.01

Activations Density 0.021%

keywords related to sports such as "ickets" and "wicket"

mentions of tickets or ticketing information

No Comments

No Known Activations