INDEX

Explanations

mentions of a person named Lore in various contexts

oai_token-act-pair · gpt-3.5-turbo

references to the name "Lore" and its variations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

akra

-0.84

ruciating

-0.79

urers

-0.75

uyomi

-0.69

insula

-0.68

inished

-0.66

ixties

-0.66

rylic

-0.65

obar

-0.65

undred

-0.65

POSITIVE LOGITS

nz

1.24

tta

1.15

tto

1.01

gon

0.96

bolt

0.95

 Lore

0.94

le

0.93

nda

0.91

suit

0.89

lette

0.88

Activations Density 0.023%

mentions of a person named Lore in various contexts

references to the name "Lore" and its variations

No Comments

No Known Activations