INDEX

Explanations

percentages expressed as words, specifically focusing on 'per cent'

oai_token-act-pair · gpt-3.5-turbo

percentages and numerical statistics in the text

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

 enthusi

-0.69

 kindly

-0.64

ModLoader

-0.64

senal

-0.62

nep

-0.60

susp

-0.60

 Raleigh

-0.58

anamo

-0.58

 realism

-0.57

ARP

-0.56

POSITIVE LOGITS

 cent

1.48

CENT

1.02

secut

0.99

missive

0.94

cent

0.90

percent

0.90

secution

0.90

iton

0.89

irteen

0.86

malink

0.85

Activations Density 0.013%

percentages expressed as words, specifically focusing on 'per cent'

percentages and numerical statistics in the text

No Comments

No Known Activations