INDEX

Explanations

references to the name "Picasso"

oai_token-act-pair · gpt-3.5-turbo

references to images and their accompanying descriptions

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

 sorely

-0.68

 prevailing

-0.59

 vigorously

-0.56

 kickoff

-0.56

pring

-0.55

 Tacoma

-0.55

 winds

-0.55

 eagerly

-0.54

mberg

-0.54

EEP

-0.54

POSITIVE LOGITS

colo

1.51

asso

1.32

nic

1.26

atin

1.07

ross

1.03

col

1.01

acho

0.99

asa

0.97

kel

0.96

ador

0.95

Activations Density 0.034%

references to the name "Picasso"

references to images and their accompanying descriptions

No Comments

No Known Activations