INDEX

Explanations

movie-related terms and titles

oai_token-act-pair · gpt-3.5-turbo

references to various types of movies and films, including genres, production details, and their cultural significance

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

bilt

-0.66

 jurisd

-0.65

esville

-0.64

ATH

-0.59

 irregularities

-0.59

emort

-0.58

fty

-0.58

aye

-0.57

lb

-0.57

 subsistence

-0.57

POSITIVE LOGITS

 starring

1.03

manship

1.01

akes

1.00

linger

1.00

 featuring

0.99

ynthesis

0.96

chool

0.96

peak

0.93

 marketed

0.91

uggest

0.88

Activations Density 0.282%

movie-related terms and titles

references to various types of movies and films, including genres, production details, and their cultural significance

No Comments

No Known Activations