INDEX

Explanations

words related to literature and novels

oai_token-act-pair · gpt-3.5-turbo

repeated references to the concept of a novel

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

tics

-0.70

cling

-0.70

adows

-0.70

ĪĴ

-0.68

 disbanded

-0.66

respect

-0.65

olid

-0.65

aviour

-0.65

adow

-0.65

henko

-0.64

POSITIVE LOGITS

 novel

1.17

ties

1.01

izations

0.93

 novels

0.90

isations

0.90

sworth

0.87

 Novel

0.83

nets

0.83

 manuscript

0.80

 screenplay

0.80

Activations Density 0.009%

words related to literature and novels

repeated references to the concept of a novel

No Comments

No Known Activations