INDEX

Explanations

names, specifically those with the sequence "na" with varying levels of specificity

oai_token-act-pair · gpt-3.5-turbo

occurrences of the substring "na" within words

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

ienced

-0.81

neys

-0.79

======

-0.71

tails

-0.71

ansas

-0.70

raved

-0.69

loo

-0.66

layer

-0.66

birds

-0.64

wolves

-0.62

POSITIVE LOGITS

eus

1.24

uthor

1.19

vel

0.92

ples

0.91

isance

0.90

ACP

0.87

Ã¯ve

0.86

emi

0.85

veland

0.81

Ã±

0.80

Activations Density 0.033%

names, specifically those with the sequence "na" with varying levels of specificity

occurrences of the substring "na" within words

No Comments

No Known Activations