INDEX

Explanations

dates that something is scheduled for

oai_token-act-pair · gpt-3.5-turbo

phrases indicating scheduled events or timings

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

ordinary

-0.72

oslav

-0.65

igans

-0.64

nex

-0.64

wrong

-0.63

é¾įå

-0.63

herty

-0.63

rats

-0.62

Guide

-0.61

 begged

-0.60

POSITIVE LOGITS

 July

1.28

 September

1.26

 August

1.25

 October

1.25

 April

1.24

 November

1.22

 June

1.21

 February

1.20

 December

1.20

 January

1.19

Activations Density 0.069%

dates that something is scheduled for

phrases indicating scheduled events or timings

No Comments

No Known Activations